Sustainable City Logistics Planning: Methods and Applications [2] 153616609X, 9781536166095

Table of contents :
Contents
Preface
Chapter 1
Electric Vehicle Charging Scheduling in Green Logistics: Challenges, Approaches and Opportunities
Abstract
Introduction
Environmental and Service Incentives
Co-Management of Public Transportations and Power Systems
EV/Charging Facility Ratio
Multiple Charge Requirements
Social Welfare
Problem Description and Challenges
Four-Element Structure of EV Charging Scheduling
Charging Requests
Resources
Constraint
Objective
Unique Features
Typical Use Cases and Problems
Vehicle Scheduling Problem with Energy Constraints
Charging Scheduling with Limited Space
Routing and Charging Station Selection
Multi-Aggregator Collaboration
Co-Management of Transportation and Power System
A Taxonomy for EV Charging Scheduling
Challenges for EV Charging Scheduling
Information Availability in Stochastic and Online Environment
Strategic Behaviors in Decentralized Environment
Existing Modelling Paradigms and Approaches
Centralized/Deterministic Modelling Paradigm
Mathematical Optimization
Meta-Heuristic Optimization
Modelling Paradigm with Uncertainties/Dynamics
Stochastic Optimization
Stochastic Programming (SP)
Robust Optimization (RO) and Distributionally Robust Optimization (DRO)
Markov Decision Process (MDP)
Online Optimization
Queueing Theory
Machine Learning Based Approach
Modelling Paradigm with Strategic Behaviors
Research Opportunities
Grid-Interactive Transportation: X + Charging Scheduling
Decentralized, Dynamic, Data-Driven: A 3D Prospective
Market-Based Mechanisms: Competition to Cooperation
References
Chapter 2
New Trends in Urban Freight Transport: How Stakeholder Engagement can Favour the Adoption of Sustainable Solutions
Abstract
Introduction
Assessing Stakeholder Preferences via Behavioural Analysis
1. Definition of the Problem
2. Preliminary Analysis
3. Survey
4. Modelling Phase
5. Scenario Simulations
Planning Policies via Living Labs
Enacting Solutions: The Case of Crowdshipping
Boosting Policy Success via Gamification
Conclusion
Acknowledgments
References
Chapter 3
Sustainable Balance between E-Commerce and In-Store Purchases
Abstract
Introduction
Sustainable Technologies to Make E-Commerce Deliveries Environmentally Friendly
Factors to Affect the Replacement of In-Store Purchases by E-Commerce
The Problem Statement and Mathematical Model Development of a City Integrated Supply System
The Impact Formalization of Integrated Supply System Functioning on the City Area
Discussion and Conclusion
References
Chapter 4
Integrating Corporate Social Responsibility in “Make or Buy” Decision: A Multi-Disciplinary and Multi-Criteria Approach
Abstract
Introduction
Literature Overview
Traditional Models
Sustainability Models
Activity Modelling by Using the Value Chain and SCOR
Value Chain Concept
SCOR/GreenSCOR Reference Model
Coupling the Value Chain with SCOR/GreenSCOR
Proposed Approach
Evaluation of the Activity’s Strategic Importance
Evaluation of the Activity’s Potential to Create a Competitive Advantage
Evaluation of the Activity’s Sustainability Performance
Illustration of the Calculation Procedure on a Hypothetical Example
Analysis of the Three Evaluations and Decision-Making
Conclusion
Acknowledgment
References
Chapter 5
Which Data to Collect and How, so as to Understand Urban Freight Distribution? Setting the Framework
Abstract
Introduction
Literature Overview
Urban Freight Distribution Indicators and Collection Methods
Data Collection Methods
Case Study Antwerp
Conclusion
Acknowledgments
References
Chapter 6
Dynamic Project Management Strategies for UK Fashion SMEs
Abstract
1. Introduction
2. Literature Review
2.1. Project Management (PM)
2.2. Project Management (PM) Success Criteria
2.3. Project Management (PM) as Management Approach
2.4. Strategy Requirements in the Fashion Industry
2.5. Dynamic Capability Requirements in the Fashion Industry
3. Research Method, Data Description and Data Analysis
3.1. Data Collection
3.2. Data Analysis
4. Findings and Discussions
4.1. PM Capabilities
4.2. External Disruptions
4.3. Internal Disruptions
4.4. Sensing New Opportunities and Linking to the Other Themes
5. Conclusion, Limitation and Future Work
6. Practical Implications
Acknowledgments
References
Chapter 7
Transportation Sustainability in Postal Industry
Abstract
Introduction
Sustainable Transportation in Postal Services
Alternative Transportation Options
Canadian Approach
Shifting towards Rail
CO2e from Air Network
CO2e from Road Network
Continuing CMB Conversions
Conclusion
References
Chapter 8
Evolution of Hub Location Problems in Sustainable Transportation Networks
Abstract
Introduction
Bibliometric Analysis of Sustainable Transportation Networks
Top Contributing Journals in the Sustainable Transportation Network
Published Studies and Their Countries of Origin
Influential Organizations Contributing to the Sustainable Transportation Network
Triple Bottom Line Framework to Analyse Key Issues in Hub Location Problems
Important Advancements in Hub Location Problems in Sustainable Transportation Networks – Economic Aspect
The Latest Arrival Hub Location Problem
The Stochastic P-Hub Center Problem with Service-Level Constraints
Design of Intermodal Logistics Networks with Hub Delays
Hierarchical Multimodal Hub Location Problem with Time-Definite Deliveries
The Design of Capacitated Intermodal Hub Networks with Different Vehicle Types
Important Advancements in Hub Location Problems in Sustainable Transportation Networks – Social Aspect
Risk Management in Uncapacitated Facility Location Models with Random Demands
Reliable Logistics Networks Design with Facility Disruptions
Solving a New Stochastic Multi-Mode P-Hub Covering Location Problem Considering Risk by a Novel Multi-Objective Algorithm
A Continuum Approximation Approach to Competitive Facility Location Design under Facility Disruption Risks
Important Advancements in Hub Location Problems in Sustainable Transportation Networks – Environmental Aspect
Sustainable Hub Location under Mixed Uncertainty
A Multi-Objective Sustainable Hub Location-Scheduling Problem for Perishable Food Supply Chain
An Interactive Possibilistic Programming Approach for a Multi-Objective Hub Location Problem: Economic and Environmental Design
Conclusion
Appendix A
References
Chapter 9
Descriptive Statistics of Problems and Solutions for the Urban Freight Transport in Brazilian Cities
Abstract
Introduction and Background
Research Approach
Area of Study
Results
Discussion
References
Chapter 10
Montreal Feedback: A Mobilization Project that Led to Concrete Solutions and the Identification of Higher Stakes
Abstract
1. Introduction
2. Context and Objectives
2.1. Difficulties of Cohabitation
2.2. Loss of Efficiency and Increased Costs
2.3. Greenhouse Gas Emissions
2.4. Pollution and Health Impacts
2.5. Congestions
3. Mobilization Process
3.1. General Approach and Lessons Learned
3.1.1. Context Favourable to Change
3.1.2. Sensitivity of the Actors and Resonance of the Findings
3.1.3. Imperatives That are Taking Over
3.1.4. Convergence on Founding Principles
3.1.5. Work on Supply but Also on Demand
3.1.6. Systemic Problem and Positive Externalities
3.1.7. Financial Assistance and Incentives
3.2. Understanding the Business Lines and Their Challenges
3.2.1. Types of Products Transported
3.2.2. Delivery Times and Punctuality
3.2.3. Return Trips
3.2.4. Parking Duration and Delivery Activities
3.2.5. Tours and Habits
3.2.6. Other Issues
3.3. Irritants
3.4. Various and Complementary Solutions
3.4.1. Express Deliveries to Individuals or Businesses
3.4.2. Delivery Impact Reduction and Simplification of Unloading in Dense Areas
3.4.3. Catering and Local Shops
3.4.4. Capacity Optimization
3.4.5. Systemic Approaches
4. Evaluation of the Leads
5. Perspectives
5.1. Consistency against Rebound Effects
5.2. A Duty of Realism
5.3. Issues that Go Beyond Cities
5.4. Taking into Account Long-Term Perspectives
Conclusion
References
About the Editor
Index
Blank Page

Citation preview

MANAGEMENT SCIENCE – THEORY AND APPLICATIONS

SUSTAINABLE CITY LOGISTICS PLANNING METHODS AND APPLICATIONS VOLUME 2

No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in rendering legal, medical or any other professional services.

MANAGEMENT SCIENCE – THEORY AND APPLICATIONS Additional books and e-books in this series can be found on Nova’s website under the Series tab.

MANAGEMENT SCIENCE – THEORY AND APPLICATIONS

SUSTAINABLE CITY LOGISTICS PLANNING METHODS AND APPLICATIONS VOLUME 2

ANJALI AWASTHI EDITOR

Copyright © 2020 by Nova Science Publishers, Inc. All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. We have partnered with Copyright Clearance Center to make it easy for you to obtain permissions to reuse content from this publication. Simply navigate to this publication’s page on Nova’s website and locate the “Get Permission” button below the title description. This button is linked directly to the title’s permission page on copyright.com. Alternatively, you can visit copyright.com and search by title, ISBN, or ISSN. For further questions about using the service on copyright.com, please contact: Copyright Clearance Center Phone: +1-(978) 750-8400 Fax: +1-(978) 750-4470 E-mail: [email protected].

NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance upon, this material. Any parts of this book based on government reports are so indicated and copyright is claimed for those parts to the extent applicable to compilations of such works. Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the Publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. Additional color graphics may be available in the e-book version of this book.

Library of Congress Cataloging-in-Publication Data ISBN: HERRN

Published by Nova Science Publishers, Inc. † New York

CONTENTS Preface Chapter 1

Chapter 2

Chapter 3

Chapter 4

vii Electric Vehicle Charging Scheduling in Green Logistics: Challenges, Approaches and Opportunities Luyang Hou, Chun Wang and Jun Yan

1

New Trends in Urban Freight Transport: How Stakeholder Engagement can Favour the Adoption of Sustainable Solutions Edoardo Marcucci, Valerio Gatta and Michela Le Pira

35

Sustainable Balance between E-Commerce and In-Store Purchases Alexander Rossolov

61

Integrating Corporate Social Responsibility in “Make or Buy” Decision: A Multi-Disciplinary and Multi-Criteria Approach Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

85

vi Chapter 5

Chapter 6

Contents Which Data to Collect and How, so as to Understand Urban Freight Distribution? Setting the Framework Katrien De Langhe, Roel Gevaers and Thierry Vanelslander Dynamic Project Management Strategies for UK Fashion SMEs Sonal Godhania, Usha Ramanathan and Ram Ramanathan

Chapter 7

Transportation Sustainability in Postal Industry Matin Foomani, Anjali Awasthi and Ciprian Alecsandru

Chapter 8

Evolution of Hub Location Problems in Sustainable Transportation Networks Ankit Sharma and Samir K. Srivastava

Chapter 9

Chapter 10

Descriptive Statistics of Problems and Solutions for the Urban Freight Transport in Brazilian Cities Leise K. Oliveira, Carla O. L. Nascimento, Luísa T. M. Sousa, Thaiza G. C. Silva, Paulo H. G. Pinto, Renata M. Pereira, Lais P. Farias, Renata L. M. Oliveira, Suellem C. Ferreira, Mylena C. R. Jesus and Iara A. M. Souza Montreal Feedback: a Mobilization Project that Led to Concrete Solutions and the Identification of Higher Stakes Mickael Brard

117

149

173

193

225

241

About the Editor

275

Index

277

Related Nova Publications

285

PREFACE City logistics planning is vital to improve goods transport in urban areas. It involves consolidation and coordination of goods transport activities to reduce the negative impacts of freight transport on city residents and their environment. It is composed of ten chapters. A brief description of the various chapters is provided as follows: Chapter 1 by Luyang Hou, Chun Wang, and Jun Yan addresses the problem of electric vehicle charging scheduling. Replacing a fossil fuelpowered car with an electric model can halve greenhouse gas emissions over the course of the vehicle’s lifetime and reduce the noise pollution in urban areas. In green logistics, a well-scheduled charging ensures an efficient operation of transportation and power systems and, at the same time, provides economical and satisfactory charging services for drivers. This paper presents a taxonomy of current electric vehicle charging scheduling problems in green logistics by analyzing its unique features with some typical use cases, such as space assignment, routing and energy management; discusses the challenges, i.e., the information availability and stakeholders’ strategic behaviors that arise in stochastic and decentralized environments; and classifies the existing approaches, as centralized, distributed and decentralized ones, that apply to these challenges. Moreover, we discuss research opportunities in applying market-based mechanisms, which shall be coordinated with stochastic optimization and

viii

Anjali Awasthi

machine learning, to the decentralized, dynamic and data-driven charging scheduling problems for the management of the future green logistics. Chapter 2 by Edoardo Marcucci, Valerio Gatta, and Michela Le Pira investigates new trends in urban freight transport and how stakeholder engagement can favour the adoption of sustainable solutions. The paper aims at contributing to the international debate on innovations in urban freight transport (UFT) by illustrating and discussing some new trends that can have the potential to revolutionize the way goods are distributed in our cities and minimize the impact UFT has on sustainability. Under this respect, stakeholder engagement becomes fundamental to foster innovation and policy adoption thus guaranteeing their spreading. A deep understanding of the behavioural issues linked to stakeholder decisionmaking is fundamental to define and implement effective policies, given the variety of heterogeneous UFT stakeholders often characterized by conflicting interests. Chapter 3 by Alexander Rossolov studies sustainable balance between e-commerce and in-store purchases. Measures for the reduction of negative influence on environment caused by freight transport in frame of ecommerce deliveries are considered. On the base of the literature analysis, the list of factors influencing on purchases execution in-stores or online was estimated. To assess the possible influence of the e-commerce growth on current cities sustainability, scenarios of online shopping development were defined. Besides, analytical models to describe the integration city transport functioning in case of trips service on the road network was described. The influence of online shopping specific weight on cities sustainability was formalized. Chapter 4 by Tasseda Boukherroub, Alain Guinet, and Julien Fondrevelle proposes a multi-criteria approach for the problem of make or buy, in the context where Corporate Social Responsibility (CSR) is considered as a source of competitive advantage for the company. The approach is inspired from the MEDIE (Modèle d’Evaluation de la Décision d’Internalisation/Externalisation) model. Potential activities to outsource (or to back-source) are evaluated against three aspects: (1) current contribution of the activity to the competitive advantage, (2) potential

Preface

ix

contribution of the activity to the competitive advantage in the future, and (3) the sustainable performance of the activity. To perform the first two evaluations, the value chain analysis and the Supply Chain Operations Reference model (SCOR) are combined. To evaluate sustainability performance, we first elect sustainability criteria and indicators based on the literature and GreenSCOR. Second, the Analytic Hierarchy Process (AHP) is used to assess the activities’ sustainability performance. Finally, the three evaluations are compared with each other and, based on the MEDIE model, different strategic decisions and options are proposed. Chapter 5 by Katrien De Langhe, Roel Gevaers, and Thierry Vanelslander sets the framework for which data to collect and how, so as to understand urban freight distribution. In order to deal with issues such as congestion and air pollution, authorities start introducing progressive measures to improve the liveability and sustainability in their urban area. However, policy makers often introduce these measures without analysing the context in which they are applied. By collecting data on a regular basis, authorities can make better decisions about their policy with respect to urban freight distribution. This study aims to contribute to this growing area of research by developing a framework of urban freight distribution indicators and data collection methods. The framework demonstrates that parts of urban freight distribution are characterised by a set of indicators and cannot be captured by measuring just one indicator. This is a very important take-away for decision makers and other urban freight stakeholders. The framework also offers the possibility to see which indicators can be measured once a certain data collection method has been chosen. This enables decision makers to collect data in a focused way, leading to time- and financial benefits. Chapter 6 by Sonal Godhania, Usha Ramanathan, and Ram Ramanathan investigates dynamic project management strategies for UK fashion SMEs. The Fashion Industry has passed through many phases, from handmade to modern technologies of manufacturing. Accordingly, the management practice in the industry has evolved from being manufacturer driven to consumer driven. In recent years, Small and Medium Enterprise (SME) businesses in this industry are facing tough

x

Anjali Awasthi

competition and challenges due to latest trends in fast fashion. Case observations, semi-structured interviews and follow up interviews are the key information source for this research. The analysis shows that projects in the fashion industry need basic understanding of Project Management (PM). By doing an in-depth case analysis, a Dynamic Project Management Capabilities (DPMC) Framework for UK fashion SMEs to improve their project management capabilities in the agile environment is suggested. The findings from the case analysis will support the interested companies in improving their dynamic project management to survive the competition. Chapter 7 by Matin Foomani, Anjali Awasthi, and Ciprian Alecsandru studies transportation sustainability in postal industry. The “last-mile deliveries” are known to be one of the biggest challenges for postal services. The impact is not only to the environment and quality of life but also the business is striving to reduce costs by deployment of various technologies and initiatives. This challenge does not limit to underdeveloped countries, emerging countries also have difficulties to establish and implement alternatives to the conventional scope of fossilbased operations. This paper proposes a method to assess alternative strategies for the last-mile of parcel deliveries, in terms of social, environmental, and economic impacts and presents an application to assess the distribution strategy. The principal of sustainability will be reviewed in respect to the nature of the postal industry, then a few examples of creativity from Canadian Postal services will be briefly presented. Finally, potentials in sustainable transportation for collection, delivery and between depots will be discussed. Chapter 8 by Ankit Sharma and Samir Srivastava presents an account of the journey of hub location problems in transportation networks from the classical single allocation p-hub median problem to the current problems involving complexities like sustainability, uncertainty, competition, environmental consciousness through government’s regulations and green sensitive consumer requirements. The purpose of this chapter is not to provide a comprehensive literature review or survey but a discussion of gradual progression and advancements in sustainable hub location problems in transportation networks context. We focus on the

Preface

xi

detailed insights covering some new turning points in location problems in the contexts of sustainability, service levels, different time zones, disruptions, multiple transportation modes, risks, etc. These insights are not only critical to the logistic firms, warehousing industry and other supply chain partners but would also play an important role in the helping research scholars to work in tandem with real-life emergent business’s requirements. Chapter 9 by Leise Kelli and Carla De Oliveira Leite Nascimento study problems and solutions for the urban freight transport in Brazilian cities. Six cities of different sizes and economic profile were analysed in the State of Minas Gerais. The data were collected in commercial establishments, and analysed through descriptive statistics and the method of successive intervals. The results obtained reflect the retailers’ point of view and indicate that there are no tendency regarding the problems and solutions. The availability of areas for loading/unloading of goods is the main problem in Betim, Belo Horizonte, Contagem and Itabira. The regulation of these areas is considered as a solution in Belo Horizonte and Itabira. The off-peak delivery is viewed as a solution in Divinópolis, Betim and Itabira. Therefore, it is necessary to investigate the reality of each city so that the policymakers could implement solutions that reduce the externalities of urban freight transport, being impossible to determine one solution as a broad and national solution to urban freight transport. Chapter 10 by Mickael Brard studies a mobilization process that led to concrete solutions and the identification of higher stakes in Montreal. In order to reduce negative externalities caused by last mile delivery in Montreal, a sample of stakeholders have been involved in a mobilization process to find some potential solutions. Above the methodology used that allowed to get to a transparent collaboration, the study offers some very concrete insights about the reality of the field. Potential solutions have to be not only looked deeper into, but provide a pragmatic base for further experimentations or analysis. Finally, as urban logistic is clearly a systemic and a long-term subject, its problems cannot be solved with a silo approach or without humility and realistically take into account future prospects.

xii

Anjali Awasthi

The book provides novel approaches to address key problems related to electric vehicle charging, stakeholder engagement, e-commerce deliveries, corporate social responsibility, urban freight data collection, dynamic project management, postal logistics, sustainable hub location, urban freight transport planning, and stakeholder mobilization. It will serve as a very useful resource for academicians and practitioners in the area.

Anjali Awasthi, Concordia University, Canada

In: Sustainable City Logistics Planning ISBN: 978-1-53616-561-6 Editor: Anjali Awasthi © 2020 Nova Science Publishers, Inc.

Chapter 1

ELECTRIC VEHICLE CHARGING SCHEDULING IN GREEN LOGISTICS: CHALLENGES, APPROACHES AND OPPORTUNITIES Luyang Hou, Chun Wang* and Jun Yan Gina Cody School of Engineering and Computer Science Concordia University Montreal, Quebec, Canada

ABSTRACT Replacing a fossil fuel-powered car with an electric model can halve greenhouse gas emissions over the course of the vehicle’s lifetime and reduce noise pollutions in urban areas. In green logistics, a wellscheduled charging ensures an efficient operation of transportation and power systems and, at the same time, provides economical and satisfactory charging services for drivers. This chapter presents a taxonomy of current electric vehicle charging scheduling problems in green logistics by analyzing its unique features with some typical use cases, such as space assignment, routing and energy management; *

Corresponding Author’s Email: [email protected].

2

Luyang Hou, Chun Wang and Jun Yan discusses the challenges, i.e., the information availability and stakeholders’ strategic behaviors that arise in stochastic and decentralized environments; and classifies the existing approaches, as centralized, distributed and decentralized ones, that apply to these challenges. Moreover, we discuss research opportunities in applying market-based mechanisms, which shall be coordinated with stochastic optimization and machine learning, to the decentralized, dynamic and data-driven charging scheduling problems for the management of future green logistics.

Keywords: green logistics, electric vehicle, charging scheduling, transportation, power system, information asymmetry/availability, environmental factors, market-based mechanism, strategic behavior, game theory, auction, machine learning, uncertainty, stochastic optimization, data-driven

INTRODUCTION Logistics has been a key sector in global economies and a crucial contributor to social progress (Juan et al. 2016, 86). Both individuals and businesses expect delivery of their goods to be faster, more flexible, and in the case of consumers, at a lower cost, which poses challenges on logistics transport. Add it all up, the city transportation sector is under intensifying pressure in delivering a better service at an ever-lower cost 1 by taxis, freights, vans, or buses. Transitively, the increasing logistics transport demands will release more greenhouse gas. Therefore, Electric Vehicles (EVs), without tailpipe emissions, are contributing to the fight against localized pollution that is increasingly important in overpopulated urban areas (Casals et al. 2016, 425-437). Governments, e.g., the UK Councils 2 , have introduced several measures to curb air pollutions in cities, such as incentivizing EVs, investing in cleaner buses, monitoring borough-wide pollutions, and pioneering the concept of low-emission zones. Some electric logistics

1 2

The future of the logistics industry- PwC, https://www.pwc.com/transport. https://techmoneyfit.com/blog/uk-government-invests-in-charging-points-for-electric-taxis/.

Electric Vehicle Charging Scheduling in Green Logistics

3

projects, such as FREVUE3 in Europe, provide insights on how innovative solutions using electric freight vehicles could help to achieve emission-free city logistics. Companies and cities are required to provide more charging service points, including central charge depot and public charging stations, for electric vehicles: decisions need to be made on the number, location, and capacity of these service points constructed. Unlike fossil fuel-powered vehicles, however, EVs must recharge frequently due to the limited driving range allowed by the battery capacities; worse still, each recharge also takes a significant amount of time. The driving range of EVs for a single charge is around one-third of the petrol-equivalent, while the recharging time can be hours, contrary to the minutes for refueling internal combustion engine vehicles (ICEVs) at a gas station (Greaves, Backman, and Ellison 2014, 226-237). For logistics, both the frequency and the duration of EV charging are concerning, especially compared to the relatively short customer service times, e.g., small package shippers or customized pickups, and thus clearly affect the productivity and route planning in logistics (Juan et al. 2016, 86).

Figure 1. EV charging scheduling in green logistics ecosystem. 3

Project: Freight Electric Vehicles in Urban Europe, https://frevue.eu/.

4

Luyang Hou, Chun Wang and Jun Yan

Given these unique features of EVs, green logistics management, on the top of traditional logistics, is emerging as a research field that aims to accommodate different types of EVs, Electric Vehicle Service Equipment (EVSE), and Renewable Energy Sources (RES). Figure 1 describes the role of EV charging scheduling played in green logistics ecosystem that charging scheduling integrates with – and deeply influence – various issues in transportation, power system, and supply chain management, etc. It is therefore highly desirable to schedule and coordinate the charging activities of EVs in city logistics in order to improve logistics efficiency and reduce operational costs, enabled by the following factors:

Environmental and Service Incentives EVs can greatly reduce the green gas emissions compared to ICEVs when it comes to running costs, environmental impacts, and quality of driving. Moreover, a well-scheduled charge can reduce drivers’ waiting time at charging stations and thus improve their satisfaction.

Co-Management of Public Transportations and Power Systems Scheduling on charging time or location will influence the service delivery of EVs and its route planning in logistics. Moreover, the load induced by EV charging at different charging stations will stress the electricity network that delivers energy to each charging station, and bring negative impacts, e.g., voltage deviation, transformer saturation, power loss, or voltage deviation (García-Villalobos et al. 2014, 717-731; Rigas, Ramchurn, and Bassiliades 2014, 1619-1635). Thus, charging scheduling should be coordinated with traffic control and energy management, in order to improve transport efficiency and maintain grid stability.

Electric Vehicle Charging Scheduling in Green Logistics

5

EV/Charging Facility Ratio Currently, the growth of publicly accessible chargers, especially fast chargers, still falls behind the increase of EVs on the road (Cazzola et al. 2016). The reason could be attributed to the large costs of charging facility investment and long payback periods. Given this, charging activities have to be well-scheduled to better utilize the limited charging capacities, which is restricted by the number of chargers installed at charging stations and its respective power output, for logistical EVs that often need to meet stringent time requirements.

Multiple Charge Requirements Logistical EVs are often tasked with heavy transport workloads at various locations throughout a day. It is thus key to schedule delivery jobs or services with multiple recharges in terms of when, where, and how much to charge during the trips. A well-designed charging scheduling can increase profits, reduce waiting time, and improve the overall efficiency of both logistics system and charging stations.

Social Welfare The efficiency of charging scheduling is highly dependent on the dynamic behaviors and decisions made by stakeholders, i.e., consumers, drivers, and charging service providers, who act as independent, rational yet mostly self-interested agents in an open decentralized environment. Thus, an overall social welfare is expected to be achieved across all stakeholders in the logistics system. Motivated by these factors, it is of great importance to efficiently coordinate and schedule different charging requests, such as single charge, deferable charge, or partial charge from users, in order to maintain charging load stability, enhance drivers and consumer experiences, as well

6

Luyang Hou, Chun Wang and Jun Yan

as improve logistics and transport efficiency. To achieve these goals, different approaches, including mathematical or stochastic optimization, game theory and machine learning, are imperative to EV charging scheduling problems in decentralized and stochastic environments so that the social welfare can be maximized in logistics systems. The remainder of this chapter is structured as follows: Section 2 reviews and analyzes some typical charging scheduling problems in sustainable city logistics and their challenges; Section 3 presents the typical approaches, including centralized, distributed, and decentralized ones, that deal with these challenges in logistical charging scheduling problems; and Section 4 provides an overview of potential research opportunities of EV charging scheduling in green logistics.

PROBLEM DESCRIPTION AND CHALLENGES As the research and development of green logistics are evolutionary, we provide a detailed summary and taxonomy of related EV charging scheduling problems. Moreover, our work complements the existing surveys by presenting the challenges for solving these charging scheduling problems in the literature.

Four-Element Structure of EV Charging Scheduling As a field in operations research, scheduling aims to find the best way to assign the limited resources to the activities at specific times such that all the constraints are satisfied, and the best objective measures are produced. Despite the variety of the definitions and models, most scheduling problems can fit in a four-element structure, which consists of resources, jobs, constraints, and objectives. The relationships of these elements are described as: resources are assigned to jobs over the

Electric Vehicle Charging Scheduling in Green Logistics

7

continuous-time or discrete-time4 and this assignment process is restricted by the constraints and is guided by the objectives. Based on this structure, we define the EV charging scheduling as a resource-constrained allocation problem in terms of the following elements:

Charging Requests A set of charging activities of EVs that is necessary for accomplishing delivery jobs or services in logistics system; Resources The space and power of charging stations. The space resource refers to the number of installed charging points; the power resources can be distributed energy generations (photovoltaic system, wind power, hydro turbines, biogas, etc.) (Yoldaş et al. 2017, 205-214), energy storage systems (ESS), and EV batteries in Vehicle-to-Grid (V2G) paradigm. Battery can also be treated as a sort of energy resource where applicable in battery swapping stations; Constraint A set of conditions that must be satisfied during charging scheduling processes, e.g., precedence constraints, release time and deadline of a charging request, battery capacity, or resource capacities. To be specific, constraints can be generally classified into three types: power capacity, limited space (parking space and charging points), and time/energy constraints of EV users; Objective A criterion to judge a schedule’s performance can be classified into two categories: from grid and charging station prospective, and from EV 4

In terms of the time representation in scheduling formulations, continuous-time models are potentially allowed to take place at any point in the continuous domain of time. While the whole optimization process in discrete-time models is split into a series of time slots and allocate energy in each time step. The mathematical programs for continuous-time problems are usually of much smaller sizes and require less computational efforts for their solutions than the discrete one.

8

Luyang Hou, Chun Wang and Jun Yan

users prospective, respectively. Charging station measures the grid stability and the utilization efficiency of its limited capacity, and users measure the quality of service (QoS), charging costs and their satisfaction on the schedules.

Unique Features Based on the four-element structure, we categorize logistical EV charging scheduling problems as a classical resource-constrained allocation problem. In contrast to the traditional scheduling problems, EV charging scheduling problems have several unique features in green logistics: 1. Service process in a market5: charging can be treated as a service process with the presence of stakeholder-provided information, and these inputs have a significant impact on the output efficiency of systems with which they interact, due to the tight coupling between stakeholders, system operation, and market exchange; 2. Transportation integration: Space assignment and routing are crucial transportation problems which decide where and when to activate the charging demands, considering drivers’ predefined deadlines, energy demands, availabilities and power limits of charging stations. The objective is to minimize the waiting time, costs, and/or travelling distances by selecting desirable locations with appropriate charging timing; 3. Energy management: Charging stations need to determine the amount of energy that can be allocated to each plug-in EV during each time period under the limitations imposed by the power distribution networks and/or energy storage systems. Energy 5

A market is one of the many varieties of systems, institutions, procedures, social relations and infrastructures whereby parties engage in exchange. It is such complex in economics that we only capture its several important concepts in conducting our research, i.e., competition, individual behaviors, price maker and taker, limited resource allocation, demand and supply, negotiation, and social welfare.

Electric Vehicle Charging Scheduling in Green Logistics

9

management is extremely important for coordinating the transport and recharge of freights or buses in the central charging depots. The energy demand for each trip, timing of each charge, as well as the capacity of each charging station should be considered in order to provide high-quality services and, at the same time, maintain the power grid stability. Moreover, charging facilities can be integrated with intermittent renewable energy sources or distributed generations, such as solar panels or wind turbines, but they may also pose uncertainties and challenges to charging scheduling; 4. Marginal utility: EV users’ gain from acquiring more energies is diminishing/decreasing along the time given the lithium battery charging profile (the current decreases at saturation stage of charging curve). In this case, pre-emption is allowed such that a user may adjust or reduce their energy demands to obtain the best time-SoC (State-of-Charge) trade-off; 5. Battery swap/switch paradigm: Battery swapping, as a new energy source instead of charging, could be an efficient and grid-friendly way for EV charging, especially for the frequent transport works in logistics. The whole swapping operation takes less than ten minutes, which is on par with conventional vehicles and even much faster than DC-fast (level-3) charge (Juan et al. 2016, 86).

Typical Use Cases and Problems In this part we analyze some typical use cases and problems of charging scheduling addressed in green logistics, based on the above features in the context of four-element structure, as follows:

Vehicle Scheduling Problem with Energy Constraints Vehicle scheduling problem models and optimizes vehicle-to-trip assignment problem with EV battery capacity constraints (Wen et al. 2016, 73-83). Each trip/task across a set of locations has a given time and energy

10

Luyang Hou, Chun Wang and Jun Yan

demand that has to be satisfied without exceeding EV’s driving range; decisions should be made on when, where and how much to charge a group of EVs while dispatching them to accomplish these trips. The objectives include to maximize the number of tasks that are completed (Rigas, Ramchurn, and Bassiliades 2015, 1339-1344), to minimize the number of vehicles used and the total distances travelled (Wen et al. 2016, 73-83), or to minimize the costs through ahead-of-time charging planning (Pelletier, Jabali, and Laporte 2018, 246-269). To achieve these objectives, it is key to estimate the energy-related cost and constraint, energy demand and time requirement of each task, under the constraints imposed by space and power capacity of the charge depots. A typical application is the scheduling of urban taxis for customer pickup (Liang, de Almeida Correia, Gonçalo Homem, and Van Arem 2016, 115-129). Taxis have to get sufficient power for the remaining driving distance of next pickup service in order to maximize the total profits, which is determined by revenues paid by passengers, vehicle maintenance costs, vehicle depreciation costs, parking space maintenance costs and/or parking costs. A well-planned charging schedule can accommodate more electric taxis in urban areas during peak hours, which can ease customer anxiety while improving driver’s profits.

Charging Scheduling with Limited Space Charging periods or start times are assigned via scheduling to EVs under time constraints (arrival, departure and charging time) and the limited number of chargers. For instance, Timpner and Wolf (2013, 579588) proposed a coordinated charging strategy to integrate the reservation and dynamic charging requests into the current charging schedules, in order to improve the utilization of the limited charging space. A genetic algorithm with tailor-made operators is proposed by García-Álvarez et al. (2018, 51-61), the objective is to minimize the total tardiness for a realworld charging scheduling problem, where the charging time is modelled as a fuzzy number. To efficiently utilize the limited charging space, a charging cable sharing strategy is proposed to deal with the public charging station

Electric Vehicle Charging Scheduling in Green Logistics

11

coordinated charging (Zhang et al. 2016, 2119-2128). The authors solved optimal configuration of charging stations and scheduling of charging power to each EV during each time interval. The objective is to enhance charging station’s utilization level and save corresponding investment costs. Similarly, a charging point sharing paradigm is proposed by Ki et al. (2018, 603-614) to balance the charging space with energy flow at a charging station. The idea is to use an M (input)-to-N (output) charger, with which the charger output and input are restricted by the limited transformer capacity.

Routing and Charging Station Selection Charging routing problem with energy constraint is to find the most economical route or charging locations with the minimum waiting/driving time, or energy consumption, taking into account the traffic conditions (path planning) and available resources at the charging stations (Gusrialdi, Qu, and Simaan 2017, 2713-2727; Yagcitekin and Uzunoglu 2016, 407419). In addition, charging station selection can be integrated with power allocation, which optimizes both transport and charging under the constraint of availability, power capacity, and price of charging stations (Iacobucci, McLellan, and Tezuka 2019, 34-52). Multi-Aggregator Collaboration In power community, load dispatching/scheduling coordinates the energy demands from multiple EVs to different charging stations in the electric power networks, in order to alleviate the negative effects of charging activities posed on electric distribution networks (GarcíaVillalobos et al. 2014, 717-731). The charging of logistical EVs at central depots should be especially well scheduled due to its heavy burdens over the local loads. Energy management is an optimal control process for the output power to the plug-in EVs during different time intervals (Sarker et al. 2017, 1127-1136; Mukherjee and Gupta 2016, 331-341), it usually does not consider the limited spaces at charging stations, but only the limited power capacity in the electric power networks. Gupta et al. (2017, 28942902) addressed a multi-aggregator-based charge scheduling problem that

12

Luyang Hou, Chun Wang and Jun Yan

incorporates collaborative charging and realistic situations with variable energy purchase and cancellation charges. The objective is to maximize the number of EVs that are scheduled at public charging stations, as well as the total profits of the aggregators.

Co-Management of Transportation and Power System A research field focuses on the systematic interaction of intelligent transportation and power system during charging scheduling, considering both of the spatial charging demands from users and the limited space and power capacity of the charging facilities (Xie et al. 2018, 817-830). This problem aims to improve the benefits of resource providers (charging stations) and resource consumers (EV users). For instance, Luo et al. (2016, 359-368) proposed a multi-objective charging scheduling strategy for EV charging scheduling and path planning, considered transport and grid related system information, such as road length, vehicle velocity, waiting time, as well as load deviation and node voltage in distribution networks. To serve more EV users with random behaviors and demands, Chen et al. (2017) proposed a two-stage stochastic programming model for charging facility planning, where charging is restricted by the limited parking space and power capacity in a multiple-charger multiple-port charging environment. A bi-level smart charging scheduling problem in working place is addressed by Yagcitekin and Uzunoglu (2016, 407-419), where the first level considers the transformer power demand and transformer capacity from the prospective of power grid; and the second level routes the EVs to the most suitable charging points, and controls the charging process cost-effectively and reliably.

A Taxonomy for EV Charging Scheduling In view of above related works, we present a taxonomy for the EV charging scheduling problems according to the operational environments, as shown in Table 1. The environmental factors are:

Electric Vehicle Charging Scheduling in Green Logistics

13

1. Information availability in an offline/deterministic, stochastic/dynamic, or online/real-time environment; 2. Information asymmetry in a centralized, distributed, or decentralized environment. The offline environment assumes all the problem data (e.g., number of jobs, charging times, release dates, due dates, charging facility information, etc.) are known in advance. As for the stochastic environment, distributions of the problem data are known in advance. While in an online environment, central controller does not know the upcoming jobs or charging requests, including number of jobs to be processed, release dates, processing times, and charging requests are presented to the controller one after another in a real-time manner. Table 1. A taxonomy of electric vehicle charging scheduling problems Offline/Deterministic Centralized

Distributed

Decentralized

Stochastic/Dynamic

(Pelletier, Jabali, and Laporte (Timpner and Wolf 2018, 246-269; Liang, de 2013, 579-588; GarcíaAlmeida Correia, Gonçalo Álvarez et al. 2018, 51Homem, and Van Arem 2016, 61; Umetani, 115-129; Liu et al. 2018, 1324; Fukushima, and Morita Wen et al. 2016, 73-83) 2017, 115-122) (Wu et al. 2017, 225-237; Wang, Wang, and Xiao 2016, 4159-4171; Zhang, Tan, and Wang 2017, 4027-4037)  (Kim, Kwak, and Chong 2017, 10111026; Xie et al. 2018, 817-830; Zhang et al. 2016, 2119-2128; Gusrialdi, Qu, and Simaan 2017, 27132727) (Liu et al. 2017, 5173-5184; (Sarker et al. 2017, Tushar et al. 2012, 1767-1778; 1127-1136; Zhu, Liu, de Hoog et al. 2015, 827-836; and Wang 2018, 2407Lam 2015, 859-869) 2419; Yoon et al. 2015, 4172-4184)

Online/ Real-time (Rigas, Ramchurn, and Bassiliades 2015, 1339-1344)



(Mukherjee and Gupta 2016, 331341; Zhang, Tan, and Wang 2017, 4027-4037)

(Gerding et al. 2011, 811-818; Parkes and Singh 2004, 791-798; Kong et al. 2016, 69-78)

14

Luyang Hou, Chun Wang and Jun Yan

In the centralized environment, central controller collects the charging requests from the relevant entities and makes decisions on how to allocate the available resources (de Hoog et al. 2015, 827-836). The centralized scheme can acquire the optimal solution, in which each entity contributes to the decision-making individually without strategic behaviors and considerations of others’ actions. The distributed and decentralized environment allow each entity to make its own decisions in a distributed environment, where entities are autonomous decision makers who are motivated by their own rationalities and not controlled by other entities or a system-wide authority. Moreover, these two schemes both assure scalability, with which scheduling related information is normally scattered across the entities in the system, and no entity has a global view of the problem. However, decentralized setting assumes that the self-interested entities may strategically misrepresent their energy consuming patterns and preferences to the resource providers, such that they can receive more economic benefits. This strategic manipulation on the outcomes may bring unexpected degradation and low efficiency to the social welfare. But the favorable point lays on entity privacy protection, which prevents the central authorities from collecting information of entities in decision making (Vardakas, Zorba, and Verikoukis 2015, 152-178).

Challenges for EV Charging Scheduling This part analyzes the complexities derived from applying current approaches to EV charging scheduling problems in green logistics, which basically arises two level of challenges: from the basic scheduling domain, and from the environment factors, respectively. The scheduling domain complexity involved in solving the most NPhard scheduling problems is the central theme of charging scheduling, which is related to the computational requirements to generate feasible outcomes given EV users’ charging requests and different sorts of constraints.

Electric Vehicle Charging Scheduling in Green Logistics

15

Aside from the computational complexity, information availability and strategic behaviors due to information asymmetry pose additional challenges on the top of traditional scheduling domain complexity in solving EV charging scheduling in logistics. These challenges will be further amplified when involving and coordinating large number of entities. Below are the details on the complexity derived from the environmental factors:

Information Availability in Stochastic and Online Environment The charging activities in green logistics are always operating in highly distributed and dynamic environments, with uncertainties coming from charging time (García-Álvarez et al. 2018, 51-61), availability of charging stations (Xie et al. 2018, 817-830), arrivals of EVs (Parkes and Singh 2004, 791-798), and energy prices (Zhang et al. 2014, 2600-2612). Different sorts of uncertainties exist in practical scenarios where some data are not precisely known. These uncertainties will influence the decisions made on when, where, and how much to charge an EV based on the target of optimization procedure. Sometimes a sequence of decisions, rather than a single decision, need to be made, e.g., taxis with several partial charges during the day trip, or highway travels with multiple charges; these decisions often depend crucially on the dynamic aspects of the environment. Moreover, uncertainties also exist in power systems, such as the state of the electricity grid, the generation of renewable sources, the charging point availability, the congestion at communication and transportation networks, and the number of EVs available to provide V2G services, which are changing quickly while a large number of EVs are either driving or charging (Rigas, Ramchurn, and Bassiliades 2014, 1619-1635). Therefore, maintaining load stability is especially challenging in Microgrids 6 due to the uncertainties from the renewable energy supplies and lower load capacity compared to the power grid.

6

A Micro-grid is a small-scale power production and delivery system comprising distributed generation facilities co-located with the loads they serve.

16

Luyang Hou, Chun Wang and Jun Yan

Strategic Behaviors in Decentralized Environment In a decentralized environment, stakeholders are not cooperative but strategic, and charging resource allocation can be viewed as a distributed optimization problem, with an objective function that depends on the strategic behaviors and private information of the stakeholders in the system. Typical examples of decentralized charging scheduling problems include task allocation across multiple self-interested shipping companies and decisions made by electric drivers on the trade-off between recharge and jobs pickup. The presence of stakeholder inputs is a necessary and sufficient condition to define a charging scheduling as a service process. Since the stakeholders are self-interested agents who aim to maximize its own utility, the challenges that significantly affect the social welfare of the whole society can be attributed to four factors from agents’ standpoint: 1. Agents may be reluctant to participate in the scheduling process; 2. Agents may misrepresent their energy demands and preferences on charging patterns, such as deadlines, charging time, or energy requirements; 3. Agents may be stubborn or insensitive to alter their charging or electricity consuming habits to gain greater benefits in response to dynamic market signals; 4. Agents may be unaware of the precise representation of their valuations or preferences. In terms of these characteristics, market-based mechanisms can be a natural way to tackle the strategic behaviors and information asymmetries in decentralized logistics systems. Typical market mechanisms refer to game theory and auction in games, which captures the conflicting economic interests of the resource providers and resource consumers (Liu et al. 2017, 5173-5184; Lam 2015, 859-869; Yoon et al. 2015, 4172-4184; Gerding et al. 2011, 811-818). Users can interact and negotiate with each other via information exchange, and further, coordinate their electricity usage with others to achieve a social welfare.

Electric Vehicle Charging Scheduling in Green Logistics

17

In addition to the incentive mechanisms in game-theoretic design, Time-of-Use (TOU), dynamic/floating pricing in demand response programs are also efficient strategies in providing incentives for EV users to alter their charging plans as their best response relied on the other users’ economic rationality, by allowing collectives of users to participate in charging scheduling (Vardakas, Zorba, and Verikoukis 2015, 152-178). The prices are retained fixed within different pricing periods ahead of time, and users receive this signal and are motivated to reduce or shift their power demands and energy usage by observing these price signals in a competitive market. However, these pricing schemes do not involve those utility theory and strategic behaviors (such as untruthful report on private information) of users in a decentralized environment, as the information revealed by users is guaranteed to be truthful. It is always challenging to elicit users’ true preferences over the charging and sensitivity on the changing of time and energy price.

EXISTING MODELLING PARADIGMS AND APPROACHES In this section, we will review the most recent activities relevant to the optimization of charging scheduling problems in logistics. Existing works typically use mathematical programming, or utility-based agent coordination combined with mechanism design, such as auction and game theory, to model the charging scheduling problems in dealing with the complexities of information availability and asymmetry. The scheduling approaches in literature can be generally classified into centralized, distributed and decentralized based approaches with respect to the aforementioned challenges. A classification of existing approaches is provided in Table 2.

Luyang Hou, Chun Wang and Jun Yan

18

Table 2. Existing approaches for the operational challenges

Centralized

Distributed

Decentralized

Offline/Deterministic Mathematical optimization; Meta-heuristics Mathematical optimization; Meta-heuristics Game theory; Combinatorial auction

Stochastic/Dynamic Robust optimization

Online/Real-time Meta-heuristics; Machine learning

Stochastic optimization; Distributionally robust optimization; Pricing strategy Auctions; Stackelberg game; Dynamic mechanism design

Online optimization/learning; Markov decision process Online mechanism design

Centralized/Deterministic Modelling Paradigm Theoretically, a centralized/deterministic modelling paradigm allows for achieving the optimal solution as the central authority has access to all information about charging scheduling. However, the difficulty of this approach lies in application bottlenecks, such as scalability, computation tractability, data privacy concerns and communication infrastructure. The model parameters (energy demand, arrival time, charging time, etc.) in centralized/deterministic environment are known with certainty even though they are only the estimations of values in real-world scenarios. Moreover, it provides optimal solutions and constitutes the basis for solving a dynamic and distributed charging scheduling problem, as witnessed by various successful applications on logistical EV charging scheduling problems (Liu et al. 2018, 1324; Pelletier, Jabali, and Laporte 2018, 246-269).

Mathematical Optimization To model and solve EV charging scheduling problems, extensive works apply Linear Programming (LP) (Umetani, Fukushima, and Morita 2017, 115-122), Dynamic Programming (DP) (Sedighizadeh, Mohammadpour, and Alavi 2019, 486-498), Mixed-Integer Linear Programming (MILP) (Liang, de Almeida Correia, Gonçalo Homem, and Van Arem 2016, 115-129; Xie et al. 2018, 817-830), decomposition

Electric Vehicle Charging Scheduling in Green Logistics

19

techniques: Lagrangian Relaxation (LR) (Zhang et al. 2014, 2600-2612), robust optimization (Sarker et al. 2017, 1127-1136; Xie et al. 2018, 817830), and stochastic programming (Wu et al. 2017, 225-237; Wang, Wang, and Xiao 2016, 4159-4171) in developing a variety sorts of scheduling algorithms or optimization strategies. These approaches follow a centralized scheme under the coordination by a central controller with known parameters. One advantageous aspect for the scheduling results is the optimal solution and highest efficiency can be obtained with the complete and truthful information from users.

Meta-Heuristic Optimization Meta-heuristic methods, such as Genetic Algorithm (GA) (GarcíaÁlvarez et al. 2018, 51-61), and Particle Swarm Optimization (PSO) (Yang, He, and Fu 2014, 82-92), etc. can efficiently explore large search space and incorporate heuristic knowledge on NP-hard problem domain. A meta-heuristic provides a sufficiently good solution to the combinatorial optimization problems, especially with incomplete/imperfect information, or limited computation capacity. Although meta-heuristics do not guarantee global optimal solutions compared to exact methods (such as simplex), they are still playing an important role in charging scheduling and path planning in a centralized environment, especially when addressing NP-hard problems with large-scale data.

Modelling Paradigm with Uncertainties/Dynamics In this part we will review some typical modelling paradigms for EV charging scheduling with uncertainties in stochastic/dynamic environments, in which the parameters are known only in probability; that is, random variables for which a probability distribution of possible parameter realizations are known, but the variability of possible values must be modelled to choose the best values for the decision variables in the optimization.

20

Luyang Hou, Chun Wang and Jun Yan

Stochastic Optimization Stochastic Programming (SP) SP is a powerful modeling framework for optimization problems under uncertainties. SP accommodates decision making process with single/two/multi-stages which optimizes the expected objective value across all the uncertainty realizations. The most widely used paradigm is two-stage model, where the deterministic optimization problem with known parameters is formulated in the first stage, and the stochastic optimization problem with unknown parameters (however its probability distribution is known) is modelled in the second stage (Zhang, Tan, and Wang 2017, 4027-4037). Specifically, SP is efficient for modelling twostage energy management in demand response programs, i.e., day-ahead demand scheduling and real-time power control, the objective is to improve the efficiency, reliability and safety of the power system, through coordinating multiple charging demands by motivating consumers to change their habits for consuming the electricity (Zhang et al. 2014, 26002612). For instance, a two-stage model by SP is proposed in (Wu et al. 2017, 225-237), where the energy scheduling with the day-ahead power market is solved in the first stage, and the real-time energy scheduling is solved in the second stage, with a goal of finding feasible solutions for all possible scenarios while minimizing the expected costs in the first stage. Similar works refer to Wang, Wang, and Xiao (2016, 4159-4171). Robust Optimization (RO) and Distributionally Robust Optimization (DRO) RO is recently introduced to model uncertainties in charging scheduling, which deals with uncertain renewable energy supplies, energy prices, and drivers’ energy demands. This approach is suitable for situations where the range of the uncertainty is known but the distribution of uncertainty is not known (Sarker et al. 2017, 1127-1136). While stochastic programming assumes there is a probabilistic description of the uncertainty, robust optimization works with a deterministic, set-based

Electric Vehicle Charging Scheduling in Green Logistics

21

description of the uncertainty, which constructs a solution that is feasible for any realization of the uncertainty in a given set. Moreover, data-driven optimization under uncertainty requires distributionally robust optimization (Sarker et al. 2017, 1127-1136; Xie et al. 2018, 817-830), also known as data-driven stochastic program, where the uncertainty is modeled by a set of probability distributions, also known as ambiguity set. DRO can obtain prior knowledge of the probability distributions through historical and/or real-time data, in terms of the practical scenarios where the precise information of the ambiguity set is rarely available or known. For instance, the day-ahead energy management model incorporated uncertain market prices using RO and used stochastic optimization to model the uncertain charging demands (arrival, departure, and charging time of EVs at charging station) (Sarker et al. 2017, 11271136). An uncertainty set is constructed for market prices to minimize the mismatch of the realized specific prices and the forecast prices, and thus may decrease charging stations’ monetary losses. Moreover, a data-driven robust optimization model is developed by Xie et al. (2018, 817-830) to optimize the capacities of renewable generations and energy storage units in each charging station, where the output uncertainty of photovoltaic energies and charging demands are formulated as robust chance constraints. Markov Decision Process (MDP) Some discrete-time stochastic charging scheduling problems can be modelled as MDP processes, with the typical time-driven scheduling policy adopted. An MDP model is typically defined as a 5-tuple: 1) decision epoch; 2) action; 3) state; 4) transition probability; and 5) reward and cost functions. MDP can investigate the constrained stochastic optimization problem in terms of the uncertainty, for instance, the uncertain EV arrivals, the intermittent renewable energy, or the variation of the energy prices (Zhang et al. 2014, 2600-2612). If the probabilities or rewards are unknown, the problem is one of reinforcement learning in practical deployment (Ko, Pack, and Leung 2018, 2165-2174; Vandael et al. 2015, 1795-1805). Typical work refers to Parkes and Singh (2004, 791-

22

Luyang Hou, Chun Wang and Jun Yan

798), which modelled the online mechanism design problem as an MDP to solve an energy unit allocation problem. The optimal policies are implemented in a truth-revealing Bayesian-Nash equilibrium.

Online Optimization Some charging scheduling problems adopt a model-free online scheme, where uncertain data are collected sequentially in a real-time fashion. The time horizon is slotted in equal intervals in time-driven mode and scheduling decisions are made at each time interval. For instance, a distributed offline and online framework is proposed by Mukherjee and Gupta (2016, 331-341) to collaborate multiple aggregators for scheduling, in order to maximize the total profits of the aggregators and the total number of vehicles that can be charged. However, Kong et al. (2016, 6978) pointed out that the major dilemma for applying the time-driven policy for charging scheduling is to determine the length of time slots: long time slot leads to few charging mode switches but causes under-utilized charging points at the stations; while short time slot improves charging point utilization but causes many mode transitions for EVs. Given this, event-driven could be an efficient solution for online charging scheduling. Queueing Theory As a typical online approach, queueing theory addresses routing and charging station selection issue in a real-time manner, in order to find the most appropriate charging locations with minimum waiting time and balance the traffic flows among different stations (Gusrialdi, Qu, and Simaan 2017, 2713-2727). This distributed scheduling method assigns multiple charges to different charging stations, which is often applied to the highway scenario. For instance, Bae and Kwasinski (2011, 394-403) proposed a spatial and temporal model to deal with EV charging demands, which first predicts the arriving rate of EVs by the fluid dynamic traffic model, and then forecasts the charging demands by queueing theory.

Electric Vehicle Charging Scheduling in Green Logistics

23

Machine Learning Based Approach Data-driven optimization under highly stochastic and distributed environments that integrates machine learning and mathematical programming is appealing in the era of big data, which can predict EV mobility, charging demands, load fluctuation, renewable energy generation, as well as other system uncertain parameters in green logistics. For example, Artificial Neural Network (ANN) is applied in the optimal energy management for a day-ahead price forecasting, so that the errors between the actual and predicted electricity prices and the costs of parking lot owner with respect to the time-of-use can be minimized (Sedighizadeh, Mohammadpour, and Alavi 2019, 486-498). Similarly, ANN with sample average approximation is used for predicting the base load power consumptions (Wu et al. 2017, 225-237). Moreover, some charging scheduling problems are modelled as MDP and are solved by reinforcement learning, which is a model-free algorithm that can directly integrate user feedback into its learning process (VázquezCanteli and Nagy 2019, 1072-1089). For instance, the cost-effective dayahead consumption plan can be learned to better forecast numerous details about each EV behavior (e.g., plug-in time, power limitation, battery size, or power curve, etc.) (Vandael et al. 2015, 1795-1805).

Modelling Paradigm with Strategic Behaviors To deal with EV users’ strategic behaviors in a decentralized environment, general equilibrium theory in game theory Economics of electric vehicle charging: game theoretic approaches (Tushar et al. 2012, 1767-1778; Yoon et al. 2015, 4172-4184) and mechanism design in microeconomic theories (Lam 2015, 859-869; Gerding et al. 2011, 811-818) incentivize users to participate in the scheduling process, report their true preferences on charging, and alter their charging or electricity consuming habits to gain greater benefits. These market-based mechanisms have gained successful applications on energy management and aggregator collaboration in an either offline (Yoon et al. 2015, 4172-4184; Liu et al.

24

Luyang Hou, Chun Wang and Jun Yan

2017, 5173-5184) or online environment (Parkes and Singh 2004, 791-798; Kong et al. 2016, 69-78). The equilibrium in game-theoretic models is defined as the condition that each player acts on her best-response strategy with respect to others’ strategy and cannot benefit herself by unilaterally deviating from this current state with an alternative strategy. For instance, the energy exchange process is modelled as a non-cooperative Stackelberg game (Tushar et al. 2012, 1767-1778), in which smart grid acts as a leader, who needs to decide on its price so as to optimize its revenue; while EV users act as followers, who need to decide on their charging strategies so as to optimize the trade-off between the benefits from battery charging and the associated charging costs. A distributed algorithm enables EVs and smart grid to reach a generalized Nash equilibrium. In addition, a cake cutting game is applied to deal with the selection of EVs and routing for transportation demands, in which the limited idle time for the serving EVs should be efficiently utilized for charging (Zhu, Liu, and Wang 2018, 2407-2419). The goal is to balance the transporting and charging demands to guarantee the long-term operations of photovoltaic systems for less charging costs and more profits. The most important mechanism-design application in a market setting is auction. In a decentralized environment, users can negotiate with the electricity network on the power allocation at different time intervals, through mechanism design-based approaches (Parkes and Singh 2004, 791-798). For example, a pricing process for multi-tenancy autonomous vehicle servicing problem is modeled as a combinatorial auction based on the Vickrey-Clarke-Groves (VCG) mechanism (Lam 2015, 859-869), where the service providers, as bidders, compete for offering transportation services; as a result, the social welfare is maximized. Moreover, de Hoog et al. (2015, 827-836) proposed a type of Groves mechanisms to allocates the available charging capacity (discrete energy unit) to users under the distribution network constraints, this mechanism is able to obtain a Nash equilibrium and is shown to be efficient and strategy-proof.

Electric Vehicle Charging Scheduling in Green Logistics

25

RESEARCH OPPORTUNITIES Multi-agent systems (MAS) architecture provides a natural modeling of the distributed and stochastic aspects of the charging markets, the existing agent-based simulation platforms from both academic and commercial sectors will provide invaluable tools for validating emerging models and techniques for solving EV charging scheduling problems in logistics. Stakeholders, e.g., electric freights/taxis/buses, charging stations, logistics companies, consumers, distribution network operators, or energy generators, can be modelled as strategic, rational and self-interested agents in the context of decentralized system engineering. In this section, we will discuss some research challenges and opportunities in designing market-based mechanisms for EV charging scheduling in decentralized and stochastic environments. Under the multiagent systems architecture, future researches are expected to go towards automated coordination systems for EV charging by applying mechanism design, game theory, stochastic optimization and machine learning based approaches, in the implementation of real-time coordination and forecasting methods that can be used by agents to adjust its forthcoming operation and properly schedule their charging.

Grid-Interactive Transportation: X + Charging Scheduling EV charging scheduling is playing a crucial role in the systematic interaction among logistics systems, intelligent transportation systems and smart grids, as its performance greatly impacts public transport efficiency, logistics costs, charging load stability, as well as agents’ satisfaction. Different charge patterns, such as single/multi-charge, deferrable charge, or partial charge with adjustable energy demands, will introduce uncertainties and flexibilities to EV charging scheduling. Add it up, charging scheduling will arise various interesting issues in the integration with highway travel/transport, ridesharing, self-driving paradigm, EV sharing, etaxi/Uber service, smart manufacturing transport, path planning,

26

Luyang Hou, Chun Wang and Jun Yan

V2G/V2B/V2V, energy management in Microgrid, and/or mobility on demand (MoD) scheme, etc. A systematic efficient outcome considering agents’ transport needs with energy demands and the limited charging capacities is supposed to be obtained, in order to maximize the systematic social welfare and resource utilization level.

Decentralized, Dynamic, Data-Driven: A 3D Prospective The interactions between stakeholders, privacy maintenance from user side, uncertainties and dynamics in charging market are key dimensions of charging scheduling problems in logistics. Dynamic information about drivers’ charging requests and charging stations’ availability have been extensively researched in dynamic and decentralized environments. For instance, in terms of an aggregated EV charging scheduling problem with energy storage, an offering/bidding strategy of an ensemble of charging stations coupled in the day-ahead electricity market is proposed by Sarker et al. (2017, 1127-1136), where aggregator determines optimal bidding strategy for the amount of energy to sell and buy from the market to meet the aggregated demands. The uncertainty modelling of the market price used robust optimization, and the aggregated charging station demand used stochastic optimization. However, big data and strategic behaviors pose additional challenges to the optimization with uncertainty. And market-based mechanism design should also be able to deal with the variations of private preferences over time. Two typical works are presented in (Feng, Narasimhan, and Parkes 2018, 354-362; Ning and You 2019): an optimal auction paradigm with deep learning is proposed by Feng, Narasimhan, and Parkes (2018, 354362), where the rules of an auction are modelled as a neural network, and machine learning is used for the automated design of auctions with budget constraints. Moreover, a closed-loop data-driven optimization framework is discussed by Ning and You (2019), where a loss function that incorporates the objective function of mathematical programming could be

Electric Vehicle Charging Scheduling in Green Logistics

27

used to train the machine learning model, and feedback from the modelbased system serves as input to the data-driven system. Several advantageous properties of model-based system compared to model-free system include strong performance guarantees and explainable outcomes. However, these techniques often do not assure scalability and may not be applicable in problems for which a priori information is unavailable (Ratliff et al. 2018). In light of this, online learning methods could be applicable, but large-scale mechanism design problems in conjunction with big data is facing with the computational challenges for training machine learning models and solving winner determination models. To our best knowledge, these is no such an effective mechanism for tackling decentralized, dynamic, and data-driven EV charging scheduling problems. Game theoretic-based auction design with machine learning in a large-scale environment, on the top of incentive mechanisms and stochastic optimization techniques, are promising research areas in green logistics.

Market-Based Mechanisms: Competition to Cooperation Game theory, stochastic optimization, and machine learning have been extremely successful in tackling a wide variety of problems in transportation systems, logistics systems and smart grids (Wang, Wang, and Xiao 2016, 4159-4171; Tushar et al. 2012, 1767-1778; Gerding et al. 2011, 811-818). Currently, considerable body of literature in game theoretic-based auction design explores competition between agents for the limited charging resources, such as energy allocation or charge reservation. Various incentive policies in a competitive market, either Groves mechanisms or dynamic pricing, encourage agents to express their true and complete preferences, or modify their habits in response to the market signals, in order to achieve a systematic efficiency. However, agents are encouraged, as not only a price-taker, but also a price-maker, to join the resource allocation process actively and pursue their benefits that are aligned with the social good. Ratliff et al. (2018)

28

Luyang Hou, Chun Wang and Jun Yan

pointed out in the presence of competition, markets could get stuck at a bad equilibrium where all agents play myopic strategies without performing sufficient exploration for a social-desirable outcome. Moreover, their truthful revelation about preferences should not be a necessity for the decision-making because they may not be able to develop a specified valuation model and the valuation may change over the course of the auction as well. In terms of this, ascending combinatorial auction could be a potential solution, where bidders can bid on single item or a package of items, discover and form their true valuations through the bidding, reveal their preferences progressively, acquire useful information by scrutinizing the bidding behaviors of their competitors, and yield the resources to moreneeded ones. As a result, it ultimately leads to an ex post equilibrium with an efficient assignment. We believe that efficient market-based mechanism design is going from competition towards cooperation among the self-interested agents, such that the information asymmetries, exogenous uncertainties from dynamic environments, endogenous uncertainties from agents’ preferences/types, and the resource constraints can be well addressed in green logistics.

REFERENCES Bae, Sungwoo. & Alexis, Kwasinski. (2011). “Spatial and Temporal Model of Electric Vehicle Charging Demand.” IEEE Transactions on Smart Grid, 3 (1), 394-403. Casals, Lluc Canals., Egoitz, Martinez-Laserna., Beatriz, Amante García. & Nerea, Nieto. (2016). “Sustainability Analysis of the Electric Vehicle use in Europe for CO2 Emissions Reduction.” Journal of Cleaner Production, 127, 425-437. Cazzola, Pierpaolo., Marine, Gorner., Schuitmaker, R. & Maroney, E. (2016). “Global EV Outlook 2016.” International Energy Agency, France.

Electric Vehicle Charging Scheduling in Green Logistics

29

Chen, Huimiao., Zechun, Hu., Haocheng, Luo., Junjie, Qin., Ram, Rajagopal. & Hongcai, Zhang. (2017). “Design and Planning of a Multiple-Charger Multiple-Port Charging System for PEV Charging Station.” IEEE Transactions on Smart Grid. de Hoog, Julian., Tansu, Alpcan., Marcus, Brazil., Doreen, Anne Thomas. & Iven, Mareels. (2015). “A Market Mechanism for Electric Vehicle Charging Under Network Constraints.” IEEE Transactions on Smart Grid, 7 (2), 827-836. Feng, Zhe., Harikrishna, Narasimhan. & David, C. Parkes. (2018). “Deep Learning for Revenue-Optimal Auctions with Budgets.” International Foundation for Autonomous Agents and Multiagent Systems. García-Álvarez, Jorge., Inés, González-Rodríguez., Camino, R. Vela., Miguel, A. González. & Sezin, Afsar. (2018). “Genetic Fuzzy Schedules for Charging Electric Vehicles.” Computers & Industrial Engineering, 121, 51-61. García-Villalobos, J., Zamora, I., San Martín, J. I., Asensio, F. J. & Aperribay, V. (2014). “Plug-in Electric Vehicles in Electric Distribution Networks: A Review of Smart Charging Approaches.” Renewable and Sustainable Energy Reviews, 38, 717-731. Gerding, Enrico H., Valentin, Robu., Sebastian, Stein., David, C. Parkes., Alex, Rogers. & Nicholas, R. Jennings. (2011). “Online Mechanism Design for Electric Vehicle Charging.” International Foundation for Autonomous Agents and Multiagent Systems. Greaves, Stephen., Henry, Backman. & Adrian, B. Ellison. (2014). “An Empirical Assessment of the Feasibility of Battery Electric Vehicles for Day-to-Day Driving.” Transportation Research Part A: Policy and Practice, 66, 226-237. Gupta, Vishu., Srikanth, Reddy Konda., Rajesh, Kumar. & Bijaya, Ketan Panigrahi. (2017). “Multiaggregator Collaborative Electric Vehicle Charge Scheduling Under Variable Energy Purchase and EV Cancelation Events.” IEEE Transactions on Industrial Informatics, 14 (7), 2894-2902. Gusrialdi, Azwirman., Zhihua, Qu. & Marwan, A. Simaan. (2017). “Distributed Scheduling and Cooperative Control for Charging of

30

Luyang Hou, Chun Wang and Jun Yan

Electric Vehicles at Highway Service Stations.” IEEE Transactions on Intelligent Transportation Systems, 18 (10), 2713-2727. Iacobucci, Riccardo., Benjamin, McLellan. & Tetsuo, Tezuka. (2019). “Optimization of Shared Autonomous Electric Vehicles Operations with Charge Scheduling and Vehicle-to-Grid.” Transportation Research Part C: Emerging Technologies, 100, 34-52. Juan, Angel., Carlos, Mendez., Javier, Faulin., Jesica, de Armas. & Scott, Grasman. (2016). “Electric Vehicles in Logistics and Transportation: A Survey on Emerging Environmental, Strategic, and Operational Challenges.” Energies, 9 (2), 86. Ki, Youngmin., Byung-In, Kim., Young, Myoung Ko., Hyemoon, Jeong. & Jeongin, Koo. (2018). “Charging Scheduling Problem of an M-to-N Electric Vehicle Charger.” Applied Mathematical Modelling, 64, 603614. Kim, Yeongjin., Jeongho, Kwak. & Song, Chong. (2017). “Dynamic Pricing, Scheduling, and Energy Management for Profit Maximization in PHEV Charging Stations.” Ieee Transactions on Vehicular Technology, 66 (2), 1011-1026. doi:10.1109/TVT.2016.2567066. Ko, Haneul., Sangheon, Pack. & Victor, C. M. Leung. (2018). “MobilityAware Vehicle-to-Grid Control Algorithm in Microgrids.” IEEE Transactions on Intelligent Transportation Systems, 19 (7), 2165-2174. Kong, Fanxin., Qiao, Xiang., Linghe, Kong. & Xue, Liu. (2016). “On-Line Event-Driven Scheduling for Electric Vehicle Charging Via Park-andCharge.” IEEE. Lam, Albert Y. S. (2015). “Combinatorial Auction-Based Pricing for Multi-Tenant Autonomous Vehicle Public Transportation System.” IEEE Transactions on Intelligent Transportation Systems, 17 (3), 859869. Liang, Xiao., de Almeida, Correia., Gonçalo, Homem. & Bart, Van Arem. (2016). “Optimizing the Service Area and Trip Selection of an Electric Automated Taxi System used for the Last Mile of Train Trips.” Transportation Research Part E: Logistics and Transportation Review, 93, 115-129.

Electric Vehicle Charging Scheduling in Green Logistics

31

Liu, Haoming., Wenqian, Yin., Xiaoling, Yuan. & Man, Niu. (2018). “Reserving Charging Decision-Making Model and Route Plan for Electric Vehicles Considering Information of Traffic and Charging Station.” Sustainability, 10 (5), 1324. Liu, Zhaoxi., Qiuwei, Wu., Shaojun, Huang., Lingfeng, Wang., Mohammad, Shahidehpour. & Yusheng, Xue. (2017). “Optimal DayAhead Charging Scheduling of Electric Vehicles through an Aggregative Game Model.” IEEE Transactions on Smart Grid, 9 (5), 5173-5184. Luo, Yugong., Tao, Zhu., Shuang, Wan., Shuwei, Zhang. & Keqiang, Li. (2016). “Optimal Charging Scheduling for Large-Scale EV (Electric Vehicle) Deployment Based on the Interaction of the Smart-Grid and Intelligent-Transport Systems.” Energy, 97, 359-368. Mukherjee, Joy Chandra. & Arobinda, Gupta. (2016). “Distributed Charge Scheduling of Plug-in Electric Vehicles using Inter-Aggregator Collaboration.” IEEE Transactions on Smart Grid, 8 (1), 331-341. Ning, Chao. & Fengqi, You. (2019). “Optimization Under Uncertainty in the Era of Big Data and Deep Learning: When Machine Learning Meets Mathematical Programming.” Computers & Chemical Engineering. Parkes, David C. & Satinder, P. Singh. (2004). “An MDP-Based Approach to Online Mechanism Design.”. Pelletier, Samuel., Ola, Jabali. & Gilbert, Laporte. (2018). “Charge Scheduling for Electric Freight Vehicles.” Transportation Research Part B: Methodological, 115, 246-269. Ratliff, Lillian J., Roy, Dong., Shreyas, Sekar. & Tanner, Fiez. (2018). “A Perspective on Incentive Design: Challenges and Opportunities.” Annual Review of Control, Robotics, and Autonomous Systems. Rigas, Emmanouil S., Sarvapali, D. Ramchurn. & Nick, Bassiliades. (2015). “Algorithms for Electric Vehicle Scheduling in Mobility-onDemand Schemes.” IEEE. Rigas, Emmanouil S., Ramchurn, Sarvapali D. & Bassiliades, Nick. (2014). “Managing Electric Vehicles in the Smart Grid using Artificial

32

Luyang Hou, Chun Wang and Jun Yan

Intelligence: A Survey.” IEEE Transactions on Intelligent Transportation Systems, 16 (4), 1619-1635. Sarker, Mushfiqur R., Hrvoje, Pandžić., Kaiwen, Sun. & Miguel, A. Ortega-Vazquez. (2017). “Optimal Operation of Aggregated Electric Vehicle Charging Stations Coupled with Energy Storage.” IET Generation, Transmission & Distribution, 12 (5), 1127-1136. Sedighizadeh, Mostafa., Amirhosein, Mohammadpour. & Seyed, Mohammad Mahdi Alavi. (2019). “A Daytime Optimal Stochastic Energy Management for EV Commercial Parking Lots by using Approximate Dynamic Programming and Hybrid Big Bang Big Crunch Algorithm.” Sustainable Cities and Society, 45, 486-498. Timpner, Julian. & Lars, Wolf. (2013). “Design and Evaluation of Charging Station Scheduling Strategies for Electric Vehicles.” IEEE Transactions on Intelligent Transportation Systems, 15 (2), 579-588. Tushar, Wayes., Walid, Saad., Vincent Poor, H. & David, B. Smith. (2012). “Economics of Electric Vehicle Charging: A Game Theoretic Approach.” IEEE Transactions on Smart Grid, 3 (4), 1767-1778. Umetani, Shunji., Yuta, Fukushima. & Hiroshi, Morita. (2017). “A Linear Programming Based Heuristic Algorithm for Charge and Discharge Scheduling of Electric Vehicles in a Building Energy Management System.” Omega-International Journal of Management Science, 67, 115-122. doi:10.1016/j.omega.2016.04.005. Vandael, Stijn., Bert, Claessens., Damien, Ernst., Tom, Holvoet. & Geert, Deconinck. (2015). “Reinforcement Learning of Heuristic EV Fleet Charging in a Day-Ahead Electricity Market.” IEEE Transactions on Smart Grid, 6 (4), 1795-1805. Vardakas, John S., Nizar, Zorba. & Christos, V. Verikoukis. (2015). “A Survey on Demand Response Programs in Smart Grids: Pricing Methods and Optimization Algorithms.” IEEE Communications Surveys & Tutorials, 17 (1), 152-178. Vázquez-Canteli, José R. & Zoltán, Nagy. (2019). “Reinforcement Learning for Demand Response: A Review of Algorithms and Modeling Techniques.” Applied Energy, 235, 1072-1089.

Electric Vehicle Charging Scheduling in Green Logistics

33

Wang, Ran., Ping, Wang. & Gaoxi, Xiao. (2016). “Two-Stage Mechanism for Massive Electric Vehicle Charging Involving Renewable Energy.” Ieee Transactions on Vehicular Technology, 65 (6), 4159-4171. doi:10.1109/TVT.2016.2523256. Wen, Min., Esben, Linde., Stefan, Ropke., Pitu, Mirchandani. & Allan, Larsen. (2016). “An Adaptive Large Neighborhood Search Heuristic for the Electric Vehicle Scheduling Problem.” Computers & Operations Research, 76, 73-83. Wu, Di., Haibo, Zeng., Chao, Lu. & Benoit, Boulet. (2017). “Two-Stage Energy Management for Office Buildings with Workplace EV Charging and Renewable Energy.” Ieee Transactions on Transportation Electrification, 3 (1), 225-237. doi:10.1109/ TTE.2017.2659626. Xie, Rui., Wei, Wei., Mohammad, E. Khodayar., Jianhui, Wang. & Shengwei, Mei. (2018). “Planning Fully Renewable Powered Charging Stations on Highways: A Data-Driven Robust Optimization Approach.” IEEE Transactions on Transportation Electrification, 4 (3), 817-830. Yagcitekin, Bunyamin. & Mehmet, Uzunoglu. (2016). “A Double-Layer Smart Charging Strategy of Electric Vehicles Taking Routing and Charge Scheduling into Account.” Applied Energy, 167, 407-419. Yang, Jun., Lifu, He. & Siyao, Fu. (2014). “An Improved PSO-Based Charging Strategy of Electric Vehicles in Electrical Distribution Grid.” Applied Energy, 128, 82-92. Yoldaş, Yeliz., Ahmet, Önen., Muyeen, S. M., Athanasios, V. Vasilakos. & İrfan, Alan. (2017). “Enhancing Smart Grid with Microgrids: Challenges and Opportunities.” Renewable and Sustainable Energy Reviews, 72, 205-214. Yoon, Sung-Guk., Young-June, Choi., Jong-Keun, Park. & Saewoong, Bahk. (2015). “Stackelberg-Game-Based Demand Response for atHome Electric Vehicle Charging.” IEEE Transactions on Vehicular Technology, 65 (6), 4172-4184. Zhang, Guanchen., Shaoqing, Tim Tan. & Gary Wang, G. (2017). “RealTime Smart Charging of Electric Vehicles for Demand Charge

34

Luyang Hou, Chun Wang and Jun Yan

Reduction at Non-Residential Sites.” IEEE Transactions on Smart Grid, 9 (5), 4027-4037. Zhang, Hongcai., Zechun, Hu., Zhiwei, Xu. & Yonghua, Song. (2016). “Optimal Planning of PEV Charging Station with Single Output Multiple Cables Charging Spots.” IEEE Transactions on Smart Grid, 8 (5), 2119-2128. Zhang, Tian., Wei, Chen., Zhu, Han. & Zhigang, Cao. (2014). “Charging Scheduling of Electric Vehicles with Local Renewable Energy Under Uncertain Electric Vehicle Arrival and Grid Power Price.” Ieee Transactions on Vehicular Technology, 63 (6), 2600-2612. doi:10.1109/TVT.2013.2295591. Zhu, Ming., Xiao-Yang, Liu. & Xiaodong, Wang. (2018). “Joint Transportation and Charging Scheduling in Public Vehicle systems— A Game Theoretic Approach.” IEEE Transactions on Intelligent Transportation Systems, 19 (8), 2407-2419.

In: Sustainable City Logistics Planning ISBN: 978-1-53616-561-6 Editor: Anjali Awasthi © 2020 Nova Science Publishers, Inc.

Chapter 2

NEW TRENDS IN URBAN FREIGHT TRANSPORT: HOW STAKEHOLDER ENGAGEMENT CAN FAVOUR THE ADOPTION OF SUSTAINABLE SOLUTIONS Edoardo Marcucci1,2, Valerio Gatta2 and Michela Le Pira3,* 1

2

Faculty of Logistics, Molde University College, Molde, Norway Department of Political Science, University of Roma Tre, Roma, Italy 3 Department of Civil Engineering and Architecture, University of Catania, Catania, Italy

ABSTRACT This paper aims at contributing to the international debate on innovations in urban freight transport (UFT) by illustrating and discussing some new trends that can have the potential to revolutionize the way goods are distributed in our cities and minimize the impact UFT has on sustainability. Under this respect, stakeholder engagement *

Corresponding Author’s Email: [email protected].

36

Edoardo Marcucci, Valerio Gatta and Michela Le Pira becomes fundamental to foster innovation and policy adoption thus guaranteeing their spreading. A deep understanding of the behavioural issues linked to stakeholder decision-making is fundamental to define and implement effective policies, given the variety of heterogeneous UFT stakeholders often characterized by conflicting interests.

Keywords: innovation in logistics, shared mobility, freight behaviour research, crowdshipping, gamification, living lab

INTRODUCTION Demographic growth and urbanization are key challenges for today’s cities. The typically overlapping peaking of passenger and freight transport demand in cities causes congestion and produces additional noxious emissions negatively affecting citizens’ health and welfare. The urban fabric should be reorganized, adapted to changing needs while an intelligent use of existing infrastructures promoted. The ever-increasing demand for better city-life standards suggests fostering a greater integration among freight activities within the urban transportation system [1]. Managing UFT requires local policy makers striking a balance between throughput, liveability, safety and sustainability. Heterogeneous stakeholders living in cities, interact, both competing and cooperating, and, often, are characterised by contrasting objectives. Besides, the traditional planning approach transport planners and practitioners follow does not consider freight a priority, focusing almost exclusively on passengers within a well-established framework [2, 3]. Under this respect, freight should be integrated in the sustainable urban mobility planning agenda [4]. In the last years, specific trends within UFT (e.g., e-commerce growth) have influenced both the nature and dimension of the challenges policy makers have to face in the near future. Innovations in the freight sector will have the potential to revoluzionize both goods management and transportation. They can be conceived as having an impact at different

New Trends in Urban Freight Transport

37

levels: technological, policy, and governance. Innovation can also play a role in promoting more sustainable behaviours [5]. Policy should address changing UFT patterns due to emerging trends like teleworking, ageing population, more densely populated urban areas, and e-commerce growth [6]. Based on this premise, this paper aims at contributing to the international debate on new trends in UFT by presenting some innovations that might substantially change the way goods are distributed in our cities and minimize the negative impact UFT has on sustainability. Under this respect, stakeholder engagement becomes fundamental to foster innovation and policy adoption while guaranteeing their acceptance, uptake and diffusion. A deep understanding of the behavioural aspects linked to stakeholder decision-making is crucial to define and implement effective policies, given the variety of heterogeneous UFT stakeholders characterised by conflicting interests. In what follows, the paper discusses four innovative concepts taking inspiration from some research carried out by the Authors in the last few years. The first two relate to methods and approaches to support policy making, the third deals with an innovative shared mobility solution, and the last refers to gamification that can foster both the adoption and spreading of the previously mentioned innovative solutions. More in detail the paper discusses: 1. Behavioural analysis – a procedure based on a modelling approach to assess stakeholder preferences. It consists of a well-thought integration of discrete choice models (DCM) with agent-based models (ABM) to account for stakeholders’ opinions in the policy making process, while mimicking their interaction to find a shared policy package [7], [8], [9]. 2. Living lab - a dynamic environment built to test project solutions in real-life contexts. Cities work as contexts for innovation and implementation processes for public and private measures cocreated with stakeholders contributing to increased efficiency and sustainable urban logistics [10].

38

Edoardo Marcucci, Valerio Gatta and Michela Le Pira 3. Crowdshipping - delivering goods via the crowd. Such an innovative solution consists of integrating passenger and freight movements to reduce the overall number of trips by increasing load factor and freight transport efficiency [11, 12]. 4. Gamification - the use of game design elements in nongame contexts to stimulate sustainable behaviours. Gamification is a promising tool that can positively influence the adoption of transport solutions. However, it should be appropriately conceived, deployed and managed to maximize users’ involvement via appropriate techniques [13].

Figure 1 shows how UFT stakeholders can be involved in the ex-ante phases of: (1) collaborative planning and evaluation of possible solutions (e.g., via living labs), taking into account their heterogeneous preferences (e.g., via a behavioural analysis), (2) implementing such solutions (e.g., crowdshipping) while fostering their success (e.g., via a gamification process).

Figure 1. New trends to foster stakeholder engagemet in UFT.

New Trends in Urban Freight Transport

39

The paper discusses the rationale behind these innovations together with recent advance in research, analysing their potential in improving UFT towards sustainability and efficiency.

ASSESSING STAKEHOLDER PREFERENCES VIA BEHAVIOURAL ANALYSIS While quantitative decision-support methods and models are wellestablished for sophisticated analysis of the transport system from an economic/technical point of view, there is a general lack of knowledge with respect to the underlying phenomena in stakeholder engagement processes [14]. Under this respect, it is important to understand how to manage different and often conflicting interests of the heterogeneous stakeholders involved. A modelling approach to study stakeholder preferences and their interacting behaviour can help to gain a deep insight of their preference structure and underlying motivations, and of the reciprocal influence with other actors. In the following, we present a procedure based on a modelling approach consisting of a well-thought-out integration of DCM with ABM as an effective way to account for stakeholders’ opinions in the policy making process. This procedure, while mimicking stakeholders’ interaction, is capable of detecting a shared and effective policy package, should it actually exist. The integration allows performing an ex-ante behavioural analysis, with the intent of testing the ex-post potential acceptability of the solutions proposed [8, 9]. Marcucci et al. [7] developed an ABM to simulate the interaction between agents (i.e., stakeholders) endowed with individual utility functions derived from DCM, in order to achieve a shared solution. Interaction in the model takes place on multiple levels (“multilayer”), reproducing a hierarchical decision process, where at the highest level there are “spokesmen” representing the interests of the categories to which they are connected in the intermediate level, while in the lower level individual category members interact with each other (see Figure 2).

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

40

“Decision” level

TOP

“Negotiation” level

MIDDLE

“Interaction” level

BOTTOM

Figure 2. Representation of the multilayer network of interaction between stakeholders [7].

The model allows simulating interaction scenarios among stakeholders who have to choose the “best” alternative within the different policy packages considered. Each stakeholder has her own utility function (derived from DCM) and interacts with her “neighbours” (i.e., stakeholders belonging to her social network) when deciding to change or maintain her opinion. For each simulated scenario, an equilibrium is reached with respect to one of the two policy packages that are pairwise compared. By simulating different scenarios, one can establish a policy ranking of packages potentially accepted by stakeholders. The key elements of the model are the following:    

Data from stated choice experiments (SCE) to estimate DCM (see e.g., [15]) Agent-specific utility functions from DCM (see e.g., [16]) Social network of relationships among stakeholders (see e.g., [17]) Opinion dynamics models to simulate stakeholder interaction (see e.g., [18]).

New Trends in Urban Freight Transport

41

The methodology was tested to reproduce the case study of UFT in Rome [7]. The analysis concerned the regulatory framework of the limited traffic zone in Rome and, in particular, the perception and preferences of the stakeholders directly involved in freight deliveries within that area. In more detail, the stakeholders involved where retailers, transport providers and own-account transport operators (i.e., retailers performing freight transport operations on their own account). Starting from a SCE survey [19, 20], latent class models [21] have been estimated and agent-specific utility functions derived, taking into account both the heterogeneity of the different categories and that of the stakeholders involved. In particular, DCM modelling results helped characterizing the ABM stakeholders and simulating different intervention policies and interaction scenarios among stakeholders. The output of the first test allowed determining a ranking of satisfactory policies that one can assume to be potentially accepted by stakeholders. Acceptability is determined in terms of a compromise between the degree of consensus and the utility perceived by the stakeholders following the interaction process [7]. Besides, a first step to validate ABM results rests on a procedure based on role-playing games (RPG) [22]. The use of RPG in ABMs is not new, and it is known as “companion modelling” approach [23]. This approach implies that stakeholders are directly involved in model definition and implementation [24]. The RPG (and the ABM) reproduce the multi-level decision-making process involving stakeholders in the definition of UFT policies. The first step is to check if the structure of the model adopted and the opinion dynamics envisioned are consistent with a real negotiation process. Subsequently, one should verify if the results derived from the ABM are in line with those derived from a first real-life experiment. Results show that the ABM developed can reproduce real-world processes and confirm that well-thought-out RPG can provide valuable insights into model performance and contribute to ABM validation. In future research, the RPG will be replicated more times with different stakeholder groups while also changing the policies under consideration. This will provide a reliable procedure for comprehensive model validation.

42

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

One can summarize the overall procedure to estimate stakeholders’ acceptability of intervention policies following the steps reported below:

1. Definition of the Problem The definition of the problem is the first important preliminary step. The issue under consideration could be of different natures, e.g., decisionmaking processes concerning (a) market-based measures, (b) regulatory measures, (c) spatial planning, (d) infrastructure measures, (e) technological measures and (f) management measures [25]. A contextual analysis is essential in this phase to clearly define the problem.

2. Preliminary Analysis The preliminary analysis aims to identify the stakeholders to involve and define the relevant policy components. In this regard, one has to select the most appropriate attributes and levels to be used in the experimental design. An efficient multi-stage multi-agent design is desirable 7 , which consists of a repeated estimate of the values of the coefficients representing the attributes, updated an appropriate number of times in order to improve the statistical significance of the estimates and/or potentially reduce the size of the sample [15].

3. Survey The survey aims at collecting the data needed to characterize the agents. In particular, SCE are useful to investigate stakeholders’ preferences for hypothetical scenarios and represent the starting point for DCM estimation. One can use qualitative questions with respect to agents’ 7

This is especially appropriate when a limited number of agents is expected to be interviewed for each category possibly included in the analysis.

New Trends in Urban Freight Transport

43

“social circle” to obtain information about the interaction between agents, and other behavioural variables.

4. Modelling Phase One estimates the DCM in the modeling phase and subsequently implement the ABM, which, in turn, will simulate interactions by using agent-specific utility functions. The interaction process is reproduced through opinion dynamics mechanisms, assuming that stakeholders decide to cooperate to find a shared decision in relation to the available alternative policies [18, 26]. Since agents are characterized by individual utility functions, they can be endowed with a willingness to change opinion, thus the likelihood of changing opinion will depend on the utility that alternative policies produce for the specific agent considered. In parallel, mathematical models could be used to simulate ex-ante the effects of transport network policies.

5. Scenario Simulations Scenario analysis allows analysing interaction dynamics between stakeholders with respect to the different policy packages and monitoring the consensus building process within the stakeholders’ network. In general, the opinion dynamics process leads to an increase of the degree of consensus as the interaction continues, while the initial utility generally decreases, as agents are willing to change their mind about the initially non-preferred option. The ex-ante evaluation of intervention policies acceptability can be combined with policies technical evaluations deriving from quantitative models/methods. The analysis performed produces an added value for the definition of intervention policies and can be framed in the general transport planning context. In fact, together with the technical and economic analyses, stakeholders’ behavioral analysis contributes to the ex-ante policy

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

44

evaluation process amenable to support the decision-makers in their choices. More details can be found in [7, 8, 9, 22]. The results deriving from this analysis, in terms of potentially accepted policies ranking, would be presented to stakeholders and policy makers for validation. This, in fact, would represent a starting point for a real participation process. Under this respect, a continuous involvement in the planning process is desirable to plan, test, and adjust solutions in such a way as described in a Living Lab approach.

PLANNING POLICIES VIA LIVING LABS Living labs (LL) are defined as dynamic environments built to test project solutions in real-life contexts: the city can typically be such a LL environment where several implementations performed by different stakeholders run in parallel [27]. The LL approach allows exploiting cities as the contexts where to test and fine tune innovative solutions identified within a given project. A city logistics LL environment could comprise three layers:   

strategic - participants interact with each other with the aim of providing a LL governance; practical - the implementations are carried out in order to get information and results about the solutions proposed; ex-post - observing the results enables a “feedback loop” so to decide about possible new LL directions [28].

In the following, the paper illustrates the LL in Rome developed within the H2020 project CITYLAB (http://www.citylab-project.eu/). The project involved partners from seven EU States, with the aim of developing knowledge and solutions to promote emission-free city logistics operations in urban centres by 2030. The project answers to the specific need of an improved knowledge and understanding of freight distribution and service

New Trends in Urban Freight Transport

45

trips and the development of best practice guidance on innovative approaches and how to replicate them8. Seven innovative solutions were tested and implemented in the cities involved using the LL approach. The LL in Rome focused on the CITYLAB intervention axis referring to urban clean waste, return trips and recycling with the aim of reducing trips by integrating direct and reverse flows (see Figure 3). The main actors involved in Rome LL were: (i) the City of Rome, (ii) Poste Italiane, the Italian national postal operator, (iii) the University of Roma Tre, providing scientific support to the LL in Rome and supervising the test bed for the implementation, (iv) Meware, a software house providing technological support for the implementation. The proposed solution in the Rome LL concerned an innovative system for integrating direct and reverse logistics flows in the urban area with the aim of improving clean waste collection, so to increase the amount of recycled materials while also minimizing the amount of transport-related CO2 emissions. As a first step of the LL, an innovative process of plastic caps collection was tested at the University of Roma Tre where the postal operator, while delivering mails, collected recyclable waste during the same transportation route by use of environmentally friendly vehicles. Before testing the proposed solution, an ex-ante behavioural analysis was performed to evaluate the degree of acceptance of the solution proposed. The analysis was based on a questionnaire administered to acquire information on stakeholders’ preferences to customize the proposed solution accordingly [10]. The first section included general information and opinions about the initiative, while the second included the SCE aimed at eliciting preferences by proposing different scenario configurations. In total, 597 interviews were administered, mostly consisting of students (90%), professors (5%) and administrative staff (5%), reflecting the different strata of daily university-going population. Three binary attributes, i.e., aim of the initiative, caps-throwing mode and 8

Horizon 2020 MOBILITY for GROWTH 2014-2015 call “Reducing impacts and costs of freight and service trips in urban areas” (H2020-MG-2014-2015).

46

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

probability to find boxes full were not significant, meaning that people were probably indifferent to the levels of that particular attribute or their preferences were not significantly affected by that factor. The significant attributes were: “Environmentally friendly transport system” and “Gamification”. Both with a positive impact on the utility function. In particular, “Gamification” was further explored and will be presented later on in this chapter. Results of the behavioural analysis were useful to plan the functioning of the proposed solution according to stakeholders’ preferences. Besides, it was estimated that in the “best” case, i.e., a scenario of plastic caps collection that would encompass all the heterogeneous preferences of the users, the potential amount of saved CO2eq per year reachable would have been of about 550 kg [10]. Under this respect, the impact of the solution proposed was high since it implied a reduction in the transport-related CO2 emissions and an increase in materials recycled. In fact, thanks to the promotional activities linked to the initiative, up to 170% more plastic caps were collected than before and an increase in the social awareness about plastic caps recycling rose from 19% to 92%.

Figure 3. Schematic representation of Rome Living Lab.

New Trends in Urban Freight Transport

47

The second cycle of the LL explored the opportunity to extend the solution to other types of recycled materials (e.g., exhausted batteries and toners) and to other geographical contexts (i.e., local, national and international) [29]. For more details, the reader can refer to Gatta et al. [10] and to CITYLAB deliverables9. It is worthy of notice that the LL approach has been adopted also for the development of the Sustainable Urban Mobility Plan of the city of Rome, in particular for the activities regarding freight transport under the responsibility of some of the Authors of this research (http://www.trelab.it/2019/04/18/logistics-living-lab-rome-sump/). In general, it would be desirable to involve stakeholders both in “plan” and “act” phases. This is especially true when investigating innovative solutions where stakeholders could play a fundamental role. This is presented in the next section.

ENACTING SOLUTIONS: THE CASE OF CROWDSHIPPING This section illustrates and discusses an innovative shared-mobility service that pools passenger and freight transport together. Integrating passenger and freight movements allows reducing the overall number of trips by increasing load factor and freight transport efficiency [11, 12]. The peculiar focus is on the necessary pre-requisites that one needs to satisfy to make crowdshipping, i.e., delivering goods via the crowd [30], successfully implemented. In fact, one can transform any trip people perform to satisfy personal needs/desires also in a freight transport service by using the available spare load capacity each individual typically has available. The idea rests on the consideration that one could stimulate a better use of currently unused transport capacity to reduce transport externalities, while performing the same amount of deliveries [31]. A typical example of crowdshipping is the “Dubbawala” service in Mumbai, India. This consists of a lunchbox delivery and return system that delivers hot lunches from homes (and restaurants) to people at work

9

Available at: http://www.citylab-project.eu/deliverables.php.

48

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

accounting for 200,000 home-cooked meals a day, corresponding to 120 tons10. Despite crowdshipping intrinsic potential, initiatives are still struggling in gaining a wider market share when it comes to UFT [30]. A critical review of urban crowdshipping initiatives offers a set of interesting suggestions with respect to possible success/failure elements. Furthermore, crowdshipping should be developed as an “environmental-friendly” service to prove socially beneficial. The best option to do so is to transform dedicated into non-dedicated trips since the least polluting trip is the one not performed. It is thus necessary to investigate delivery models that can make use of commuters’ trips that would be performed anyhow so to avoid generating additional ones [32]. It is appropriate to focus on commuters’ trips since they are typically frequent and predictable. Recent research focuses on the following three main aspects: (1) supply, i.e., under which conditions passengers would be willing to act as crowdshippers ([11, 12, 32]); (2) demand, i.e., under which conditions people would be willing to receive their goods via a crowdshipping service ([11, 12, 32]); (3) measurement of the potential impacts/outcomes this solution might have from an economic and environmental perspective [33]. Advanced techniques to study the potential adoption of this innovative solutions were used (i.e., SCE and DCM) allowing for a sound analysis of different crowdshipping future scenarios and estimating their main associated impacts. Finally, in order to account more realistically for last-mile delivery operations, a hybrid dynamic traffic simulation is adopted such that the macroscopic features of traffic (triggering of congestion, queue spillbacks and interactions with traffic signals) could be reproduced in combination with the microscopic features of delivery operations (delivery vehicles are tracked along their routes) [34]. The above mentioned papers refer to a case study performed in the city of Rome, characterized by a population of around 3 million people performing approximally 700,000 thousand trips during the morning peak,

10

https://www.telegraph.co.uk/travel/cruises/articles/delivery-service-mumbai-dabbawalas/.

New Trends in Urban Freight Transport

49

and employing roughly 32,700 vehicles a day to perform more than 35,000 loading/unloading operations in the city centre [12]. In the first step of the research, Marcucci et al. [11] investigated the main characteristics of an innovative crowdshipping initiative in Rome. Pursuing this aim, they administered a questionnaire to approximately 200 students enrolled at University of Roma Tre. Students can be considered, in general, “early adopters/providers”. The main results indicate that 87% of the interviewees would be willing to act as crowdshippers, while 93% would accept to receive goods through a crowdshipping system. In general, no one judges crowdshipping as a “bad idea”. With respect to the supply side, like for other crowdshipping services around the world, the most important condition for effective cooperation is economic. Apart from this, interviewees want to be sure that they do not dedicate too much time to transport the goods. This implies that the deviation from their normal path must be minimized. However, on average, they are prepared to deviate a maximum of 24 km. This distance fits the average distance of present urban crowdshipping initiatives of 8-30 km. No one of the interviewees would make more than 5 intermediate stops, and almost 80% of students prefer to do only 1 or 2 intermediate stops. A last aspect to be considered is that crowdshippers want to preserve their privacy and, therefore, they are in general unwilling to be traced (57%). Based on the results of this research, Serafini et al. [32] used SCE to identify the most important features associated with the choice of acting as a crowdshipper and DCM to study the underlying behaviour. The SCE refer to the city of Rome and its metro network assuming deliveries in a B2C market. The paper assumes packages can be picked-up/dropped-off in Automated Parcel Lokers (APL) located either inside metro stations or in their surroundings. Data were collected by administering 240 interviews to metro users in October 2017. APL location proved to be the attribute with the greatest impact, while delivery booking the one with the smallest relevance. Having APL inside the metro stations (instead of outside) was considered more important than the remuneration level (considering the range used in the

50

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

survey was 1 – 3€/delivery). Real-time booking was preferred with respect to the off-line option. However, this characteristic was considered less important than others suggesting that people need to get organized to produce a crowdshipping service using public transport. Gatta et al. [12] reported the SCE survey results investigating crowdshipping demand. The survey explores the demand-side and investigates the role service time, service cost, parcel tracking availability (available/not-available), delivery schedule date/time flexibility (available/non-available) play in stimulating potential e-commerce users to choose a crowdshipping service when receiving goods bought online. Most crowdshipping service potential users declared to prefer to pick-up the parcel during the afternoon (38%) or evening (33%) and to have the pickup option available for 24 h (44%). Only 9% of the respondents declared to prefer having a short pick-up time (less than 3 h), mainly due to safety issues. The possibility to plan delivery date and timing constitutes the most relevant feature, while a shorter shipping time with respect to the present situation has the lowest impact on utility. This reflects the fact that the present delivery system is, in general, efficient in terms of shipping time (e.g., same-day delivery), while time windows are usually wide, and people have to wait at home their goods, engendering either dissatisfaction or provoking missed deliveries. Starting from the results obtained, Gatta et al. [33] provided an estimation of the environmental and economic impacts a public transportbased crowdshipping service might have in urban areas. Several scenarios with different features of the service are proposed and evaluated up to 2025 in terms of both externalities (local and global pollutant emissions, noise emissions and accidents reductions) and revenues. Finally, Simoni et al. [34] studied the potential impacts on traffic and pollution deriving from the implementation of alternative crowdshipping practices in Rome by means of a simulation approach. Externalities were investigated at the network level by analyzing the effects of the implementation of such service in comparison to traditional delivery framework for parcels. The analyses showed that, depending on different implementation features, crowdshipping could result in very different

New Trends in Urban Freight Transport

51

changes in emissions and traffic congestion and that the chosen transportation mode by crowdshippers plays the main role in identifying the effects of this solution. Under this respect, crowdsourced deliveries by car have generally higher negative impacts than corresponding deliveries by public transit. All these results help understanding and quantifying the potential of this strategy for last mile B2C deliveries. Moreover, they provide local policy makers and freight companies with a good knowledge base for the future development of a green crowdshipping platform and for estimating the likely impact the system could have both from an economic and environmental point of view. In order to guarantee the success of such solutions and influence behaviours, elements typical of games can be used, as will be illustrated in the next section.

BOOSTING POLICY SUCCESS VIA GAMIFICATION “Gamification” is a promising tool that can positively influence the adoption of transport solutions. It takes advantage of the power of game mechanics for non-entertainment purpose to foster behaviour change towards sustainability [35]. Behavior change is the end goal policy makers aim for. In fact, a voluntary change in behaviour can contribute to the substantial changes needed to ensure a sustainable society [36]. Inducing behaviour changes in the freight industry could help jointly achieve significant reductions in the externalities produced as well as improve economic productivity and efficiency. For this reason, gamification is progressively used both in in passenger (e.g., [37, 38, 39]) and freight (e.g., [40, 41]) transport policies. However, gamification is not capable per se to induce behaviour change. One should rather appropriately conceive, deploy and manage it to maximize users’ involvement via appropriate techniques. Marcucci et al. [13] propose an innovative method to ensure the fundamental prerequisites necessary to reliably build a full-fledged user-

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

52

centered gamification process. More in detail, they illustrate, through a specific gamification example in reverse logistics, how one could acquire through SCE the necessary information and knowledge to appropriately structure, at a strategic level, a gamification process capable of producing the desired engagement and behavior change results. In fact, it would be explicitly based upon the preferences of the given “players’ types” present in the target population. The case study illustrated refers to Rome LL within CITYLAB previously described. It concentrates on developing a user-centered gamification process to stimulate engagement and participation in the plastic cap recycling initiative. In line with other studies [42], preliminary analysis conducted testifies to the high potential a gamified recycling initiative has thanks to its engaging capability within a university environment. The paper provides a theoretically based and easy-to-implement approach to maximize the potential success of gamification developed, so to ensure the maximum beneficial impacts of the innovations proposed. It explores gamification design from a strategic level, leaving the implementation phase to a second step of the research. However, it can be easily replicated both at tactical and operational level. Figure 4 reports the schematic representation of the proposed procedure based on [13]: 

 

Given the gamification objective, SCE, accounting for playertypes and game elements/mechanics, helps tailoring the gamification process based on the preferences the stakeholders expressed. In the implementation phase, players’ behavior respond to game dynamics. Game monitoring will allow checking the correspondence between envisaged player-types and actual ones, and possibly adjusting game elements/mechanics.

New Trends in Urban Freight Transport

53

Figure 4. Schematic representation of user-centered gamification design [13].

Game designers have to specify goals, rules, settings, contexts, types of interactions and boundaries of the situation to be gamified [43]. The fundamental game elements and mechanics are: 1) point assignment mechanism (e.g., by overcoming levels, succeeding in a mission), 2) rewarding mechanisms (e.g., based on badge, external reward such as discounts), and 3) type of participation (e.g., individual, team). SCE are used to tailor, at a strategic level, the most important game components and align them with agents’ preferences/expectations so to maximize each agent-type engagement/behaviour change. The attributes used in the survey were: 1) rewarding system (RS), 2) point assignment mechanism (PAM) and 3) participation type (PT). The levels used in the SCE were: RS: (a) internal (i.e., badge) or (b) external (i.e., discount) to the game; PAM: (a) succeeding in a mission, (b) making a single “virtuous” action or (c) competing with other players; PT: (a) individual, (b) in self-defined teams, (c) in teams consisting of entire departments, (d) in teams consisting of different categories (i.e., students, professors and technical/administrative staff), (e) hybrid (i.e., both individual and in team).

54

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

“Type of Participation” (51%) was the most relevant attribute followed by “Rewarding” (35%) and “Point assignment” (14%). Results also showed that: (i) external rewards were preferred to internal ones; (ii) mission was the most preferred mechanism for point assignment, while competition (i.e., winning against other participants) was the least preferred; (iii) individual participation was the least preferred option, while hybrid participation the most preferred, where users play individually, but at the same time act as a team. The main finding of the paper relates to the new approach adopted for gamification design with respect to current literature. In general, the issue of gamification design is not addressed, and the basic approach to tackle it suggests identifying player-types and qualitatively deriving the most suitable game mechanics for the specific player-type setting. Practical implications for policy makers relate to the possibility to use the proposed procedure and replicate it whenever a gamification approach is suitable to foster behavior change in the transport sector.

CONCLUSION This chapter discussed four innovative concepts related to UFT taking inspiration from some research carried out by the Authors in the last few years. In particular, it focused on methods and approaches to support policy making (a behavioural analysis of stakeholder preferences and the living lab approach), innovative shared mobility solutions (i.e., crowdshipping), and the role of gamification to foster both the adoption and spreading of the previously mentioned innovative solutions. All these topics were addressed via well-known and robust methods, which rest on behavioural modelling theory, i.e., SCE, DCM and ABM. The rationale is to study behavioural aspects linked to stakeholder decision-making and to understand how to possibly influence behaviours for sustainable and effective policies, given the variety of heterogeneous UFT stakeholders. Under this respect, new trends in the urban freight sector, often connected with the spreading of new techonologies, will influence both the

New Trends in Urban Freight Transport

55

nature and dimension of the challenges policy makers have to face in the near future. Stakeholder engagement becomes a fundamental component to foster innovation and policy adoption while guaranteeing their acceptance, uptake and diffusion. Future research should further address the problem of innovative policy making, accounting for (1) the potential of new technology, (2) stakeholder participation both in planning and in implementation, and (3) stakeholder cooperation.

ACKNOWLEDGMENTS Both Edoardo Marcucci and Valerio Gatta would like to acknowledge the EU financial support for the CITYLAB project financed within the H2020 framework (grant agreement no. 635898).

REFERENCES [1]

[2]

[3] [4]

Gatta, V., Marcucci, E. & Le Pira, M. (2017). Smart urban freight planning process: integrating desk, living lab and modelling approaches in decision-making. European Transport Research Review, 9(3), 32. Lindholm, M. (2013). Urban freight transport from a local authority perspective–a literature review. European Transport\Trasporti Europei, Issue 54, Paper n° 3, ISSN 1825-3997. de Dios Ortuzar, J. & Willumsen, L. G. (2004). Modelling transport. John Wiley & Sons. Wefering, F., Rupprecht, S., Bührmann, S. & Böhler-Baedeker, S. (2013). Guidelines. developing and implementing a sustainable urban mobility plan. Available at: https://www.eltis.org/sites/default/files/ guidelines-developing-and-implementing-a-sump_final_web_ jan2014b.pdf.

56 [5]

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

EC - European Commission. (2011). White Paper. Roadmap to a Single European Transport Area – Towards a competitive and resource efficient transport system. COM(2011) 144 final. [6] ALICE/ERTRAC (2015). Urban Freight research roadmap. Available at: http:// www.ertrac.org/ uploads/ documentsearch/ id36/ ERTRAC_Alice_Urban_Freight.pdf. [7] Marcucci, E., Le Pira, M., Gatta, V., Inturri, G., Ignaccolo, M. & Pluchino, A. (2017). Simulating participatory urban freight transport policy-making: Accounting for heterogeneous stakeholders’ preferences and interaction effects. Transportation Research Part E: Logistics and Transportation Review, 103, 69-86. [8] Le Pira, M., Marcucci, E., Gatta, V., Inturri, G., Ignaccolo, M. & Pluchino, A. (2017). Integrating discrete choice models and agentbased models for ex-ante evaluation of stakeholder policy acceptability in urban freight transport. Research in transportation economics, 64, 13-25. [9] Le Pira, M., Marcucci, E., Gatta, V., Ignaccolo, M., Inturri, G. & Pluchino, A. (2017). Towards a decision-support procedure to foster stakeholder involvement and acceptability of urban freight transport policies. European Transport Research Review, 9(4), 54. [10] Gatta, V., Marcucci, E., Le Pira, M. & Ciccorelli, A. (2018). Integrating direct and reverse logistics in a “living lab” context: evaluating stakeholder acceptability and the potential of gamification to foster sustainable urban freight transport. In City Logistics 3. Towards Sustainable and Liveable Cities, (pp. 1-18). Wiley. [11] Marcucci, E., Le Pira, M., Carrocci, C. S., Gatta, V. & Pieralice, E. (2017). Connected shared mobility for passengers and freight: Investigating the potential of crowdshipping in urban areas. In 2017 5th IEEE International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS), (pp. 839-843). IEEE. [12] Gatta, V., Marcucci, E., Nigro, M. & Serafini, S. (2019). Sustainable urban freight transport adopting public transport-based crowdshipping for B2C deliveries. European Transport Research Review, 11(1), 13.

New Trends in Urban Freight Transport

57

[13] Marcucci, E., Gatta, V. & Le Pira, M. (2018). Gamification design to foster stakeholder engagement and behavior change: An application to urban freight transport. Transportation Research Part A: Policy and Practice, 118, 119-132. [14] Le Pira, M. (2018). Transport Planning with Stakeholders: An AgentBased Modeling Approach. International Journal of Transport Economics, 45(1). [15] Gatta, V. & Marcucci, E. (2016). Stakeholder-specific data acquisition and urban freight policy evaluation: evidence, implications and new suggestions. Transport Reviews, 36(5), 585609. [16] Gatta, V. & Marcucci, E. (2014). Urban freight transport and policy changes: Improving decision makers’ awareness via an agent-specific approach. Transport policy, 36, 248-252. [17] Le Pira, M., Inturri, G., Ignaccolo, M. & Pluchino, A. (2018). Dealing with the complexity of stakeholder interaction in participatory transport planning. In Advanced Concepts, Methodologies and Technologies for Transportation and Logistics, (pp. 54-72). Springer, Cham. [18] Le Pira, M., Inturri, G., Ignaccolo, M., Pluchino, A. & Rapisarda, A. (2017). Finding shared decisions in stakeholder networks: an agentbased approach. Physica A: Statistical Mechanics and its Applications, 466, 277-287. [19] Marcucci, E., Gatta, V. & Scaccia, L. (2015). Urban freight, parking and pricing policies: An evaluation from a transport providers’ perspective. Transportation Research Part A: Policy and Practice, 74, 239-249. [20] Gatta, V., Marcucci, E., Le Pira, M., Scaccia, L. & Delle Site, P. (2018). Willingness to pay measures to tailor policies and foster stakeholder acceptability in urban freight transport. Scienze Regionali, 17(3), 351-370. [21] Boxall, P. C. & Adamowicz, W. (2002). Understanding heterogeneous preferences in random utility models: A latent class approach. Environmental and Resource Economics, 23, pp. 421-446.

58

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

[22] Le Pira, M., Marcucci, E. & Gatta, V. (2017). Role-playing games as a mean to validate agent-based models: An application to stakeholder-driven urban freight transport policy-making. Transportation Research Procedia, 27, 404-411. [23] Voinov, A. & Bousquet, F. (2010). Modelling with stakeholders. Environmental Modelling & Software, 25(11), 1268-1281. [24] Barreteau, O., Le Page, C. & D’aquino, P. (2003). Role-playing games, models and negotiation processes. Journal of Artificial Societies and Social Simulation, 6(2). [25] Marcucci, E. & Gatta, V. (2017). Investigating the potential for offhour deliveries in the city of Rome: Retailers’ perceptions and stated reactions. Transportation Research Part A: Policy and Practice, 102, 142-156. [26] Castellano, C., Fortunato, S. & Loreto, V. (2009). Statistical physics of social dynamics. Reviews of modern physics, 81(2), 591. [27] CITYLAB. (2016). CITYLAB local living lab roadmaps. H2020 CITYLAB Project, Deliverable 3.2. [28] CITYLAB. (2016). Practical guidelines for establishing and running a city logistics living lab, H2020 CITYLAB Project, Deliverable 3.1. [29] CITYLAB. (2017). Sustainability analysis of the CITYLAB solutions. H2020 CITYLAB Project, Deliverable 5.4. [30] McKinnon, A. C. (2016). Crowdshipping. A communal approach to reducing urban traffic levels? Logistics White Paper 1/2016. DOI: 10.13140/RG.2.2.20271.53925. [31] Bubner, N., Helbig, R. & Jeske, M. (2014). Logistics trend radar. DHL trend research. http://www.supplychain247.com/paper/ 2014_logistics_trend_radar_delivering_insight_today_creating_value _tomorrow. [32] Serafini, S., Nigro, M., Gatta, V. & Marcucci, E. (2018). Sustainable crowdshipping using public transport: a case study evaluation in Rome. Transportation Research Procedia, 30, 101-110. [33] Gatta, V., Marcucci, E., Nigro, M., Patella, S. & Serafini, S. (2019). Public Transport-Based Crowdshipping for Sustainable City

New Trends in Urban Freight Transport

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

59

Logistics: Assessing Economic and Environmental Impacts. Sustainability, 11(1), 145. Simoni, M. D., Marcucci, E., Gatta, V., & Claudel, C. G. (2019). Potential last-mile impacts of crowdshipping services: a simulationbased evaluation. Transportation, 1-22. https://doi.org/10.1007/ s11116-019-10028-4 Nelson, M. J. (2012). Soviet and American precursors to the gamification of work. Proceedings of the 16th International Academic MindTrek Conference. Presented at MindTrek’, 12, ACM, 2012, 23–26. Schwanen, T., Banister, D. & Anable, J. (2012). Rethinking habits and their role in behaviour change: the case of low-carbon mobility. Journal of Transport Geography, 24, 522-532. Meloni, I. & di Teulada, B. S. (2015). I-Pet Individual Persuasive Eco-travel Technology: A tool for VTBC program implementation. Transportation Research Procedia, 11, 422-433. Hoh, B., Yan, T., Ganesan, D., Tracton, K., Iwuchukwu, T. & Lee, J. S. (2012). Trucentive: A game-theoretic incentive platform for trustworthy mobile crowdsourcing parking services. In 2012 15th International IEEE Conference on Intelligent Transportation Systems, (pp. 160-166). IEEE. Kazhamiakin, R., Marconi, A., Perillo, M., Pistore, M., Valetto, G., Piras, L., Avesani, F. & Perri, N. (2015). Using gamification to incentivize sustainable urban mobility. In 2015 IEEE First International Smart Cities Conference (ISC2), (pp. 1-6). IEEE. Klemke, R., Kravcik, M. & Bohuschke, F. (2013). Energy-efficient and safe driving using a situation-aware gamification approach in logistics. In International Conference on Games and Learning Alliance, (pp. 3-15). Springer, Cham. Hense, J., Klevers, M., Sailer, M., Horenburg, T., Mandl, H. & Günthner, W. (2013). Using gamification to enhance staff motivation in logistics. In International Simulation and Gaming Association Conference, (pp. 206-213). Springer, Cham.

60

Edoardo Marcucci, Valerio Gatta and Michela Le Pira

[42] Berengueres, J., Alsuwairi, F., Zaki, N. & Ng, T. (2013). Gamification of a recycle bin with emoticons. In 2013 8th ACM/IEEE International Conference on Human-Robot Interaction (HRI), (pp. 83-84). IEEE. [43] Robson, K., Plangger, K., Kietzmann, J. H., McCarthy, I. & Pitt, L. (2015). Is it all a game? Understanding the principles of gamification. Business Horizons, 58, 411-420.

In: Sustainable City Logistics Planning ISBN: 978-1-53616-561-6 Editor: Anjali Awasthi © 2020 Nova Science Publishers, Inc.

Chapter 3

SUSTAINABLE BALANCE BETWEEN E-COMMERCE AND IN-STORE PURCHASES Alexander Rossolov* Transport Systems and Logistics Department O. M. Beketov National University of Urban Economy in Kharkiv Kharkiv, Ukraine

ABSTRACT Currently e-commerce has become preferable for most people by making the purchasing process easy and less time-consuming. Availability of home delivery services allows to minimize the time spent in the physical shops, if the end consumer prefers it. But in the case of urbanization trend, e-commerce growth may cause some problems, especially for the environment. Due to this, the aim of the chapter is to reveal the current measures for creating the sustainable effect in case of goods deliveries in urban area under e-commerce growth conditions. The chapter deals with the estimation results for the factors influencing on instores or online purchases from the current trends and research reviews analysis within the sustainable city logistics. To assess the possible influence of e-commerce growth on current urban sustainibility the scenarios of online shopping development have been presented. The *

Corresponding Author’s Email: [email protected].

62

Alexander Rossolov chapter reveals the analytical models to formalize the functioning of integrated urban transport network in case to serve the trips of different purposes. The influence of online shopping specific weight on urban sustainibility has been given through mathematical formulation.

Keywords: e-commerce, sustainability, home delivery, carbon dioxide

INTRODUCTION The high pace of life in present megalopolises implies the active utilization of the city infrastructure by population. The size of commodity bundle supposes the consumption of big list of goods and foodstuffs that can be purchased by two ways: in the physical stores or in the Internet shops. In the face of time pressure, people aim to minimize the time expenditures for non-production activity such as purchases. Under such conditions the presence of the Internet shops with home delivery services become more popular and widely used each day. According to statistics in the USA in 2017 the number of online shoppers was of 215.4 million people with the positive dynamics of this number change. By 2021, it is expected that the number of customers in the Internet (US residents) will reach 230.5 million people. The e-commerce development causes the list of changes of social and economic character. Thus, the obvious advantage of the Internet shopping is the release of people’s time that they previously spent on shopping trips in the physical stores. This saved time people may spend on selfdevelopment or other creative activity. As the positive economic aftermath it can be noticed the decrease of final price of goods and the reduction of demand for retail space and staff serving the purchases. However, it should be noted that the purchases on-line with home delivery services fulfilment may cause the negative aftermath. So, creating the extra freight flows in frame of B2C, the e-commerce predestinates the increase of freight traffic value. This leads to a list of negative impacts as increase of CO2 gases value, cars flow and public transit speed rate

Sustainable Balance between E-Commerce and In-Store Purchases 63 reduction, the deficit growth of parking space and sometimes provoke of jams because of double parking problem [1]. When the home delivery service for e-commerce has been started, the assumption that the number of shopping trips in the physical stores should reduce has been expected to become true. But in fact, this service generated only the decrease of purchasing volume in the physical stores without significant changes in the number of trips [2]. In addition, the increase in the freight vehicles’ mileage was detected in case of the “pickup points” delivery service. This service may push people to shift from public transit to private cars in making the pick up delivery more comfortable [3]. Thus, the possitive and negative effects of e-commerce suppose the determination of some sustainable conditions of it functioning. It means that the total amount of goods for an urban territory should be serviced in specific ratio between the online and in-store purchases to provide the sustainable development of the city. Through this, the negative impact reduction from e-commerce and the strengthening of the positive effects from its usage can be reached.

SUSTAINABLE TECHNOLOGIES TO MAKE E-COMMERCE DELIVERIES ENVIRONMENTALLY FRIENDLY In last ten years the intensive development of e-commerce has been leading to growth of the service users all around the world. By 2017 about 83% of population of China made the purchases online and countries of Western Europe and the USA approximately were at this level too (80% and 77% accordingly) [4]. The research of Holguín-Veras and Cleckley (2016) allowed determining the average number of the Internet supplies in a day per one person in New York metropolitan area. According to the research results, the value of online supplies has grown from 0.054 deliveries per person-day (2014) to 0.10 deliveries per person-day (2016) in case of New York City. And for such high dynamics, the domination of B2C system supplies compared with B2B deliveries within the

64

Alexander Rossolov

metropolitan areas may become relevant. As e-commerce growth catalysts, the home delivery service can be considered to promote the buyers in reducing the shopping activity time. But the home delivery is less consolidated that may cause the increasing in goods final price [6]. More over, in the case of attended delivery, the end consumer absence at home may cause the iterated supplies to make the transport cost bigger. Because of this, the pick-up points arrangement has become widely used avoiding the need to personally contact with the end customer. So, the new delivery service makes it possible to disclose the negative consequences generated by e-commerce. As a means of solving these problems of traffic congestion and CO2 gases volume reduction, the off-hour delivery (OHD) can be used [7]. In the case of New York City, the international group of scientists has substantiated and put into practise the delivery system of goods to the end consumers in the time window from 7 pm to 6 am. This initiative is a striking example of sustainable development measures and can be implemented under the conditions of coordinated action presence among the local authorities (representing the public sector), the private sector (the serviced stores and the market areas) and the academia (a group of scientists providing a theoretical substantiation and methodology of OHD). The OHD implementation for New York City has allowed to reduce the transportation cost from $3.07 per km (regular hours deliveries) to $1.99 per km (OHD). Considering the parking fines, the transportation cost decrease has been reached up to 45%. The second positive OHD effect is the reduction of СО2 emissions to 60% for New York City. Because of the OHD introduction, the sustainable effect will be the higher, if the population of the metropolitan area becomes bigger [8]. Currently there are several countries that are planning to reduce the amount of СО2 gases by implementation of the OHD, namely: China, United States, Japan, India, South Korea, Brazil, Indonesia, Pakistan, Mexico and Italy [8]. In this list, it can be pointed the countries with the active ecommerce growth (United States, Japan, South Korea, Italy and China). So, the OHD introduction will allow taking the sustainable effect of B2C supplies with the usage of the pick-up points.

Sustainable Balance between E-Commerce and In-Store Purchases 65 As an alternative option to solve the traffic density issue is the use of the goods consolidation paradigm. For this purpose, the supply chains of multi-echelon type can be introduced with distribution centres and local depots [9, 10]. They allow cutting the transportation cost, but for the last mile logistics their effectiveness may be not high because of the road network overload and the supply demand stochasticity being the main feature of the Internet purchases. Furthermore, the stationary position of local depots and distribution centres makes the supply chain system less flexible, if the service demand is of high variability. The effective way to solve this problem is the usage of mobile depots as the distribution centres on a small area [11]. It makes possible to implement the strategy of goods consolidated distribution to any part of the urban area. Such technology is realized by TNT Express in Brussels via the two echelon-based last mile parcel deliveries. The first echelon produces the consolidated delivery to the inner city by trailer holding on board of the electric vehicles. When the truck reaches the city centre, it is unloaded, and the last mile logistics is realized by ecofriendly electrically supported tricycles. The involved tricycles are more maneuverable comparing with standard light trucks and generate zero СО2 emissions. Thus, the mobile depots technology forms the sustainable effect and is economically effective even for the regular hours deliveries. It can be implemented for the Internet deliveries within the last mile logistics to reduce the negative impact of home deliveries, if e-commerce activity volume will grow. On a par with mobile depots to achieve the sustainable effect for ecommerce deliveries the two echelon systems with cargo bikes on the second echelon is widely used [12, 13]. These systems may be created according to the classic principles of stationary local depots [12] and with the usage of mobile reloading points [13]. Under the stationary reloading points usage, the main task for effective two echelon system functioning is determination of the depot rational location. During the design stage, the major aspect is the determination of alternative possible positions for local depot along the inner area perimeter. This type of urban supply system (USS) construction reflects one of the city authorities measure to form the

66

Alexander Rossolov

sustainable effect of city development, namely, nearby delivery area (NDA) system [16]. As the demand for e-commerce deliveries may be stochastic for some goods, the depot location problem has become not trivial. One of the ways to solve this task, proposed by Naumov and Starczewski (2019), is the algorithm to simulate the potential freight turnover of the cargo bikes on the second echelon under the condition of the depot position different scenarios. The simulation results to draw the conclusion that the freight turnover of cargo bikes under stochastic demand is the normally distributed random variable. The rational depot location has to be found according to the scenario to generate the minimum freight turnover of cargo bikes [12]. Under the variable demand conditions Anderluh et al. (2017) proposed to use the hybrid two echelon system to suppose the vans as the vehicles and depots within the second echelon. The authors have offered the algorithm for determination of mobile depot location to reload the goods on cargo bikes synchronously with VRP solving. The difference in speed and vehicles’ carrying capacities supposes the time frame synchronization to fulfil the reload and end consumers’ deliveries. The additional condition of this delivery system is the service area segmentation in relation to the inner city to be serviced by ecology friendly cargo bikes. Thus, the implementation of the OHD and multi-echelon systems with ecofriendly vehicles usage on the final echelon may generate the double effect. The first one is to reduce the carbon dioxide volume and the second one is to decrease of transportation cost being very important for goods deliveries according to the last mile logistics [14, 15].

FACTORS TO AFFECT THE REPLACEMENT OF IN-STORE PURCHASES BY E-COMMERCE From day to day the number of people that are ready to buy in the Internet becomes more. When e-commerce was on the stage of genesis, the books and audio CDs have been of the biggest specific weight of sales via the Internet. Nowadays the list of products that are available in the Internet

Sustainable Balance between E-Commerce and In-Store Purchases 67 shops is equal to goods range from the physical stores [1, 4]. People may buy online both clothes and hygiene products till building materials with home delivery service. Certainly, with the current market prices, the consumer tends to buy the goods at lower prices willing to purchase mainly in the Internet shops. It is due to the goods in the physical stores are offered at higher prices as usually. But not always it plays a major role. According to [17] the statement that the less book shops or clothes stores are situated in the dwelling area the more people from this area tend to buy online. So, the passive accessibility of stores according to T. Sinai и J. Waldfogel (2004) plays an important role in formation of consumer preferences between the Internet and physical stores. At the same time, the research made by Cao et al. (2013) allowed to determine the influence of people education level on buying method choice. By analyzing the purchases in Minneapolis-St. Paul metropolitan area Cao et al. (2013) found that people from the inner-city households tend to buy mostly online despite the high accessibility of physical stores in the dwelling area. Moreover, under conditions of low store accessibility the people from suburban area prefer to purchase in the physical stores but not in the Internet. The key factor to form people preferences in purchase method according to [18] is the education level of agglomeration district inhabitants. So, as usually in the city centre live people with higher education level then people from the suburban area that is why the downtown inhabitants are better skilled in the Internet services [18]. In the research made by Russo and Comi (2012), the elements of the utility function for purchases in the physical stores for a list of goods are revealed. As the significant factors to form the utility function, the authors have proposed to use the number of employees at retailer related to freight type s in the zone d, the distance between zone o and d and the dummy variable which equals to 1 for intrazonal trips and 0 otherwise. Also, the authors pointed out that the mileage between zone o and d makes the inversely proportional influence on utility function value. Thus, the probability of making the purchase will be decreased under conditions of the distance growth to the physical store [21]. However, the research results by Russo and Comi (2012) also expose the sharp increase in

68

Alexander Rossolov

purchases dimension if the purchase travel is being done to non-easily accessible zones. In such way the authors have emphasised that passive accessibility plays the significant role in formation of purchase dimension. The list of the factors to influence the purchasing behaviour may be enlarged with economic factors such as the household or one-person income. Weltevreden et al. (2007) have revealed the straight correlation between the income level and a car ownership in the household in course of buying activity. The authors have determined that for the car owners the stores accessibility is insignificant at all in time of purchase trips. They prefer to purchase in the Internet. While for people traveling on public transport, the store accessibility is of a significant impact on the choice of purchase type between the Internet and the physical stores. An important aspect in assessing of the factors influence on purchases mode choice may be the monitoring of the sales amount increase in the online shops. For example, when the end consumer prefers to buy online with home delivery service it may entail to increase in the above-stated negative negative environmental effects the ability to forecast the online shopping rate will allow to estimate the possible environmental damage for e-commerce development could cause and to generate the necessary measures for decreasing it. According to this, Comi and Nuzzolo in [22] have described and simulated the possible scenarios of the in-store and online purchases under e-commerce growth conditions. The scenario to simulate the e-shopping attitudes increase (according to statistics of Italy, every year online shopping growth equals to 18%) has revealed the unproportional less quantity of possible purchases in the physical stores comparing withecommerce growth. The biggest decrease has been specified for electronics (-11.7%) and category “other goods” in amount of -7.30%. In such conditions it is obviously that the number of the Internet deliveries will grow. Thus, this will cause the rise of cars and transit trips in the physical stores too according to [20]. Zhai et al. (2017) indicated that as usual the Internet purchases provoked the trial trips in the physical stores. In this case the online shopping demand growth may lead to unproportional increase in the transit and private cars trips making additional pressure on

Sustainable Balance between E-Commerce and In-Store Purchases 69 road network and public transit. In turn, it will form the negative environmental impact and the urban area sustainability level. Under such conditions, the possible states of the city transport system (transit, supply chain and road network) in the case of different specific weight of the online purchases in general goods flow in the metropolitan have to be studied.

THE PROBLEM STATEMENT AND MATHEMATICAL MODEL DEVELOPMENT OF A CITY INTEGRATED SUPPLY SYSTEM Nowadays, in the metropolitan areas the most of the deliveries for restocking or for servicing the end consumers are fulfiled by the freight transport (supply chain) being integrated into the city network. Under the conditions of online shopping increase, the value of trips and vehicles on the road network will grow too causing changes in general integrated transport system. The results of the theoretical formalization of the Internet purсhases (namely, its specific weight in the city total trade amount) influence on sustainability of the city are presented on the Figure 1. According to Figure 1 both the low and the high specific weight of ecommerce delivery will cause the reduction of the city sustainability. The current society is on the verge of e-commerce to become the major type of purchases. This will transform all trade system and its infrastructure. To estimate all possible scenarios of e-commerce growth influence on the city area, at first, the general model of the supply chain operating in time of its integration into a city transport network has to be developed. So, the formal composition of supply chain created from the integrated city delivery system is proposed to consider as the structure of following components:

DS  G, , ,  ,

(1)

70

Alexander Rossolov

where DS is the model of the end consumers material supply; G is the road network to implement the transportation process;  is the storageretrieval system providing the supply of small shops, malls, post offices, Internet shops or home delivery;  is the urban transit network operating in the city;  is the demand for goods supply service. The G road network formalization is originated from its classical components from the graph theory [23]:

G  ,  ,

(2)

where  is the finite set of vertices (nodes);  is the set of edges.

Figure 1. The formal influence of Internet purchases specific weight on city sustainability.

The vertex position is determined according to two-dimensional coordinate system:    i  xi , yi , i  1; R  , j  1; E  ,

(3)

Sustainable Balance between E-Commerce and In-Store Purchases 71 where  i is the vertex of the graph; x, y is the position coordinates of i vertex in two-dimensional model; R is the number of the vertices in the graph, unit; E is the number of the edges in the graph model, unit. The second subsystem of the road network model has the numerical characteristics. They are the base for the  modelling:





   j  l j , n j ,i ,i 1  ,

(4)

where  j is the edge of the graph; l j is the length of j edge in the graph, km; n j is the number of traffic lanes for j edge in the graph, unit;  i is the vertex of the edge beginning;  i 1 is the vertex of the edge ending. The storage-retrieval system accomplishes a function of store, distribution and goods delivery via urban transport network. According to functional features of the storage-retrieval system, it may be marked out three components to provide the full cycle of the goods supply:

  , ,  ,

(5)

where  is a subsystem of the freight transport vehicles;  is the set of warehouse infrastructure elements;  is the set of retail shops on the city territory.

  v1 , v2 ,..., vk  ,

(6)

where v is a type of the freight vehicle to be described by the body type and loading capacity; k is the number of freight vehicles type, unit. 

    g

f

e

,

(7)

where g is the distribution warehouses in the service area;  f is the post offices and collection points in the service area;  s is the warehouses of

72

Alexander Rossolov

the Internet shops for self-pickup service; g is the number of distribution warehouses in the service area, unit; f is a number of post offices and collection points on the city territory, unit; e is a cumulative number of the Internet shops warehouses on the city territory, unit. g    , S ,

(8)

where   is the transport node (vertex of the graph) of distribution warehouse location; S is the square of the distribution warehouse, metres2. The models  f and  s are formed by analogy to the model g . 

m ,h  q

d

,

(9)

where mq is the big shopping centres of the “mall” and “supermarket” type; hd is the stores of small commercial areas (for example, the “neighbourhood store” type); q is the number of big shopping centres in the service area, unit; d is the number of the stores of small commercial areas, unit. The models mq and hd have the similar structure with the model g and are determined according to the graph vertexes (spatial parameter) and area capacity (surface parameter). The obligatory condition of this research is the urban transit network inclusion to the model of the urban material supply system. It is substantiated by the possibility of the urban transit as the link between the storage-retrieval system and the end-consumers sector. In time of the urban transit operation, the different types of the trips such as “home-to-work,” “work-to-store,” “store-to-home,” “study-to-store” and etc. may be serviced [24], which have the specific time frames and the volume features. So, during research of the USS it must be taken into account that public transit may service the pairs of trips with the “store” element (for example, “work-to-store,” “study-to-store” and etc.) making the alternative to home delivery. That is why the system of public transit has to be

Sustainable Balance between E-Commerce and In-Store Purchases 73 included into integrated urban supply chain. The model of urban transit network may be formalized as following:

  ,  ,

(10)

where  is the set of transit routes;  is the demand for goods purchase realized on city transit network.

  r1 , r2 ,..., rR  ,

(11)

where r is the route of city transit; R is the number of city transit routes, unit. r  so , bc  ,

(12)

where so is the stops servicing the specific transit route; bc is the rolling stock operating on the route. The major feature of the stops is its spatial characteristic that in the model frame is reflected by the graph’s vertices (nodes). Each stop refers to some vertex of the graph. In addition, the stop may service several modes of the transit (bus, trolley bus) or only one of them. According to this, the second parameter of every stop is mode of public transit that it may service. From this, the model of the stops so is of the following:

so   s , ,

(13)

where  s is the transport node (a vertex of graph) where the public transit stop is located;  is the mode of public transit that may be serviced by the stop. The set bc describes the rolling stock subsystem. The main task of this subsystem is to service the trip demand without its failure. It means that the capacity and quantity of the rolling stock play a critical role in the service

74

Alexander Rossolov

process. Following the functional nature of the rolling stock, the model that describes this subsystem can be presented as:

bc  vb , nb ,

(14)

where vb is the passenger transport vehicle capacity, pass.; nb is the number of transit vehicles operating on the route, unit. The purchase activity via public transit may have several sources of its origin. Some of people prefer to buy the goods in the physical stores after work. In this case, we obtain the pair of the “work-to-physical store” trips. Also, this pair of trips may be origin in the case of not-labour activity, for example, fitness, generating the “gym-to-physical store” pair. But the public transit may be used for e-commerce trips too, for example, for such pairs of trips as “work-to-post office,” “work-to-parcel lockers” and etc. According to different types of the trips pairs a demand for goods purchase using public transit is proposed to formalize as the set of travels:

 11y ...1yj ...1yU     ...................      iy1 ... ijy ... iUy  ,   0, y  1;Y  ,  ...................   y  y   U 1...Ujy ...UU  

(15)

where  is the number of travels between i and j transport zones, pass.; U is the number of transport zones in the model of city transit network, unit; y is the travel type, for example “home-store”; Y is the general number of travel types in case of purchase activity, unit. And the final element of DS model to reflect the demand for goods supply in the urban area has the following structure:

  , ,  ,

(16)

Sustainable Balance between E-Commerce and In-Store Purchases 75 where  is the set of matrices describing the supply links between consignors and consignees;  is the set of matrices of consignments between consignors and consignees on the network;  is the set of delivery time matrices.

  a1 , a2 ,..., a  ,

(17)

where a is the matrix of the supply links between the supply chain players within the city territory; λ is the number of the pair links for material supply between the players within the city territory (Figure 2). Consignors

Consignees Distribution Internetwarehouse store ware(centre) house

Mall (supermarket)

Store of small shopping capacity

Post office

End consumer (home delivery)

Distribution warehouse (centre) Internet-store warehouse Mall (supermarket) Store of small shopping capacity Post office End consumer (home delivery)

Figure 2. The matrix of the goods supply channels.

The dark blue colour reflects the pair links to implement the supply functions. The end of this supply chain is the end consumer. In the model, the grey colour indicates the links in the supply chain that can be used in the case of the reverse deliveries. These reverse material flows imply the failure in the supply process and have the list of negative effects [25]. The operation of the reverse supply channel emerges in the case of home

76

Alexander Rossolov

delivery when the end consumer is out of the household. This process has the stochastic nature. The absence of the links in the integrated supply chain is marked with the white colour on Figure 2. So, the total demand for goods supply for entire urban network is proposed to formalize as the set of demand layers:

  1 , 2 ,...,   ,

(18)

where  is the matrix of the supplies in frame of the one pair link, for example, “distribution centre-to-the store of small shopping capacity”; ε is the number of the demand layers for goods deliveries between different players in the urban supply chain, unit. The structured model of total demand service with the integrated supply chain is shown on Figure 3.

Figure 3. The diagram of supply channels to the end consumer.

The delivery time plays an important role in city logistics. The inaccurate prediction of the supply time causes the failure of the supply process and leads to costs increase. As well, in the case of home delivery the inaccurate prediction of the delivery time is the major factor of revers supply to arise. It produces the extra vehicle kilometres and, as the

Sustainable Balance between E-Commerce and In-Store Purchases 77 aftermath, generates the negative environment impact [10]. In frame of instore delivery the prediction of the supply time must be executed with considering of city policies implemented, for example, time windows [16] causing the significant changes in vehicles routes [26, 27]. In the case of sustainable policies in cities, another measure is the “Off-Hours Deliveries Project” [7, 8] that require the significant changes of the delivery time (from day time deliveries to night supplies). So, all these measures must be taken into account in frame of the integrated city supply system creation. From the above mentioned the subsystem  is of the following form:

  t1 , t2 ,..., t  ,

(19)

where  is the number of interaction channels involved in goods supply (Figure 2). So, each channel of interaction will be characterized by the specific number of consignors D and consignee A. In turn, the delivery times model is formalized as matrix:  11 ... 1j ... 1A     ...................  t    i1 ...  ij ...  iA  ,   1;   ,  ...................          D1...  Dj ...  DA  

(20)

where  is the time point of goods delivery, h:mm.

THE IMPACT FORMALIZATION OF INTEGRATED SUPPLY SYSTEM FUNCTIONING ON THE CITY AREA The main target of the integrated supply system functioning is delivery of goods to the end consumers in the city area. This process consists of store, sorting and transportation subprocesses. As the major negative effect

78

Alexander Rossolov

of these operations, the environment pollution has to be pointed out. So, grounding to above-mentioned as the general impact of the supply system operation, the following function is presented: GCDS  WCDS  TCDS  ECDS ,

(21)

where GCDS is the total costs related with integrated supply system functioning; WCDS is the store and handling goods costs in the integrated supply system; TC DS is the transportation costs for the goods delivery in the supply channels (both freight vehicle and city transit service of purchase activity); ECDS is the assessment of the transportation negative environment impact (СО2 emissions, noise pollution). WCDS   DC   i   PC   j  ig

j f

  IC  k   MC  m h   SC  h l ke

hq

,

(22)

ld

where DC    is the goods store costs in the distribution centre; PC   is the goods reception, post offices store and collection points costs; IC   is the goods reception and the internet-shops warehouse store costs; MC  m  is the store and goods placement in malls and supermarkets costs; SC  h  is the store and goods placement in stores of small shopping capacity costs. TCDS   FCDS ( da )   PCDS (ijz )   CCDS (ijz ) , d D aA 

iU jU zY

(23)

iU jU zY

where FCDS is the goods transportation costs via the supply channels; PCDS is the city transit costs raised in the case of the of the end consumers

purchase activity; CCDS is the private car trip costs in the case of purchase activity.

Sustainable Balance between E-Commerce and In-Store Purchases 79 ECDS   FEDS ( da )   PEDS (ijz )   CEDS (ijz ) , d D aA 

iU jU zY

(24)

iU jU zY

where FE DS is the size of the environmental damage as the result of goods transportation in the supply chain; PE DS is the size of the environmental damage as the result of city transit purchase trips; CE DS is the size of the environmental damage as a result of private car trips in the case of purchase activity. The total costs related to integrated supply system functioning will depend on volume of the goods and its shares among the supply channels. It can be illustrated by the following formula: Qtotal   Qtotal   i ,

(25)

i

where Qtotal is the general value of the goods that has been serviced in the supply system per a time period, ton;  is the share of goods to be serviced by i supply channel. The set of total  shares can be transformed into two types deliveries: in-store and e-commerce supplies. In this case the model (25) will be transformed to: Qtotal  Qtotal    Qtotal  1    ,

(26)

where  is the average specific weight of the total amount of goods realized via e-commerce services (Figure 1). The service of the Qtotal   and Qtotal  (1   ) values will cause the different results of the integrated supply system functioning ( GCDS ). In this case, the variability of GCDS allows to obtain the optimal (rational) state of the USS that will reduce the negative general impact of its operation. One of the ways to solve this problem is to study the impact of  variability on GCDS using the models developed.

80

Alexander Rossolov

DISCUSSION AND CONCLUSION The analytical research revealed that current megalopolises have been faced with problems of transport systems to be overloaded by private cars and freight vans. This essentially deteriorates the environmental component of the urban area to reduce the city’s liveability. At the same time, the reduction of the delivery system efficiency provokes significant rise of transportation prime cost within the last mile logistics. Time expenditures have become of high relevancy because of cities’ inhabitants high transport activity. The rise of home delivery services results in the vehicles’ extra mileage because of less freight consolidation, iterated supplies and trial trips on using private or public transport. All these negative aspects must be levelled in the nearest future along with the implementation of the sustainable urban development strategies. The chapter presents the major aspects of cities sustainability improvement under e-commerce increase. The measures to level the negative impact because of the e-shopping goods delivery have been determined. Among the key efforts, the off-hour deliveries, mobile depots and creation of two echelon systems with the cargo bikes to operate within the second echelon have been proposed. Under assessment of factors to affect the “from in-stores to online shopping” shift, the significance of people education level, variability in trip distances and passive accessibility of physical stores have been outlined. As the result of this analytical study, the formalization of the eshopping specific weight influence on the city sustainability has been shown. It has been determined that under condition of e-commerce growth, the number of accompanying (trial) trips forming to create the additional congestion in the urban transport network would increase. To estimate this process, the mathematical model has been developed. In course of further research from the model obtained it is necessary to simulate the functioning of the integrated supply system with aim of determining the different e-purchasing specific weight in total goods for the urban area.

Sustainable Balance between E-Commerce and In-Store Purchases 81

REFERENCES [1]

[2]

[3] [4]

[5]

[6] [7]

[8]

Francke, J., and Visser, J. (2015). Internet shopping and its impacts on mobility, KiM Netherlands Institute for Transport Policy Analysis, 14. Visser, J., Nemoto, T., and Browne, M. (2014). Home Delivery and the Impacts on Urban Freight Transport: A Review. Procedia Social and Behavioral Sciences 125:15-27. BESTUFS. 2001. E-commerce and urban freight distribution (home shopping). Best Practice Handbook Year 2, Germany: PTV. (2001). Maat, K. and Konings, R. (2018). Accessibility or Innovation? Store Shopping Trips versus Online Shopping. Transportation Research Record 2672 (50):1-10. Holguín-Veras, J., and Cleckley, E. (2016). Use of freight trip generation techniques to manage curb space. Webinar recording; moderator: Jeff Wojtowicz. Accessed January 1, 2017, https://coesufs.org/wordpress/ peer-to-peer-exchange-program/webinar16/. TNO. (2010). Online winkelen in Nederland. Noodzaak tot slimmere logistiek, TNO, Delft. Holguín-Veras, J., Hodge, S., Wojtowicz, J., Singh, C., Wang, C., Jaller, M., Aros-Vera, F., Ozbay, K., Weeks, A., Replogle, M., Ukegbu, C., Ban, J., Brom, M., Campbell, S., Sanchez-Díaz, I., González-Calderón, C., Kornhauser, A., Simon, M., McSherry, S., Rahman, A., Encarnación, T., Yang, X., Ramírez-Ríos, D., Kalahashti, L., Amaya, J., Silas, M., Allen, B., and Cruz, B. (2018). The New York City Off-Hour Delivery Program: A Business and Community-Friendly Sustainability Program. Interfaces 48.1:70-86. Holguín-Veras, J., Encarnación, T., González-Calderón, C. A., Winebrake, J., Wang, C., Kyle, S., Herazo-Padilla, N., Kalahasthi, L., Adarme, W., Cantillo, V., Yoshizaki, H. and Garrido, R. (2018). Direct Impacts of Off-Hour Deliveries on Urban Freight Emissions. Transportation Research Part D Transport and Environment 61: 84103.

82 [9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

Alexander Rossolov Nuzzolo, A., Persia, L. and Polimeni, A. (2018). Agent-Based Simulation of urban goods distribution: a literature review. Transportation Research Procedia 30: 33-42. Goodchild, A., Wygonik, E. and Mayes, N. (2018). An analytical model for vehicle miles traveled and carbon emissions for goods delivery scenarios. European Transport Research Review 10(8): 10. doi.org/10.1007/s12544-017-0280-6. Verlinde, S., and Macharis, C. (2016). Innovation in urban freight transport: the Triple Helix model. Transportation Research Procedia 14: 1250-1259. Naumov, V. and Starczewski, J. (2019). Approach to simulations of goods deliveries with the use of cargo bicycles. AIP Conference Proceedings 2078:5. doi.org/10.1063/1.5092073. Anderluh, A., Hemmelmayr, V. C. and Nolz, P. C. (2017). Synchronizing vans and cargo bikes in a city distribution network. Central European Journal of Operations Research 25(2): 345-376. Rodrigue, J.-P., Dablanc, L. and Giuliano, G. (2017). The Freight Landscape: Convergence and Divergence in Urban Freight Distribution. Journal of Transport and Land Use 11(1): 1-16. Rossolov, A., Popova, N., Kopytkov, D., Rossolova, H. and Zaporozhtseva, H. (2018). Assessing the impact of parameters for the last mile logistics system on creation of the added value of goods. Eastern-European Journal of Enterprise Technologies 5/3(95): 7079. Russo, F. and Comi, A. (2018). From City Logistics Theories to City Logistics Planning. In City Logistics 3: Towards Sustainable and Liveable Cities, edited by Eiichi Taniguchi and Russell G. Thompson, 329-348. doi:10.1002/9781119425472. Sinai, T. and Waldfogel, J. (2004). Geography and the Internet: Is the Internet a Substitute or a Complement for Cities? Journal of Urban Economics 56(1): 1-24. Cao, X. J., Chen, Q. and Choo, S. (2013). Geographic Distribution of E-shopping; Application of Structural Equation Models in the Twin

Sustainable Balance between E-Commerce and In-Store Purchases 83

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

Cities of Minnesota. Transportation Research Record: Journal of the Transportation Research Board 2383: 18-26. Weltevreden, J. W. J. and Van Rietbergen, T. (2007). E-shopping Versus City Centre Shopping: The Role of Perceived City Centre Attractiveness. Tijdschrift voor Economische en Sociale Geografie 98(1): 68-85. Zhai, Q., Cao, J., Mokhtarian, P. L. and Zhen, F. (2017). The Interactions between E-Shopping and Store Shopping in the Shopping Process for Search Goods and Experience Goods. Transportation 44(5): 885-904. Russo, F. and Comi, A. (2012). The simulation of shopping trips at urban scale: Attraction macro-model. Procedia  Social and Behavioral Sciences 39: 387-399. Comi, A. and Nuzzolo, A. (2016). Exploring the relationships between e-shopping attitudes and urban freight transport. Transportation Research Procedia 12: 399-412. Huang, Y., Verbraeck, A. and Seck, M. D. (2016). Graph Transformation Based Simulation Model Generation. Journal of Simulation 10(4): 283-309. Vrtic, M., Lohse, D., Fröhlich, P., Schiller, C., Schüssler, N., Teichert, H. and Axhausen, K. W. (2007). A simultaneous twodimensionally constraint disaggregate trip generation, distribution and mode choice model: Theory and application for the Swiss national model. Transportation Research Part A: Policy and Practice 41(9): 857-873. Nuzzolo, A., Comi, A. and Papa, E. (2014). Simulating the effects of shopping attitudes on urban goods distribution. Procedia – Social and Behavioral Sciences 111: 370-379. Gevaers, R., Van de Voorde, E. and Vanelslander, T. (2011). Characteristics and typology of last-mile logistics from an innovation perspective in an urban context. In City distribution and Urban freight transport: Multiple perspectives, edited by Cathy Macharis and Sandra Melo, Edward Elgar Publishing 56-71.

84

Alexander Rossolov

[27] Kok, A. L., Meyer, C. M., Kopfer, H. and Schutten, J. M. J. (2010). A Dynamic Programming Heuristic for the Vehicle Routing Problem with Time Windows and the European Community Social Legislation. Transportation Science 44: 442-454.

In: Sustainable City Logistics Planning ISBN: 978-1-53616-561-6 Editor: Anjali Awasthi © 2020 Nova Science Publishers, Inc.

Chapter 4

INTEGRATING CORPORATE SOCIAL RESPONSIBILITY IN “MAKE OR BUY” DECISION: A MULTI-DISCIPLINARY AND MULTI-CRITERIA APPROACH Tasseda Boukherroub1,2,3,*, Alain Guinet4 and Julien Fondrevelle4 1

Department of Systems Engineering & Numerix Laboratory, Ecole de technologie supérieure, Montreal, QC, Canada 2 Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation (CIRRELT), University of Montreal, Montreal, QC, Canada 3 FORAC research Consortium, Université Laval, Québec, QC, Canada 4 Laboratoire Décision et Information pour les Systèmes de Production (DISP), INSA-Lyon, Villeurbanne, France

ABSTRACT This study proposes a multi-criteria approach for the problem of make or buy, in the context where Corporate Social Responsibility (CSR) *

Corresponding Author’s Email: [email protected].

86

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle is considered as a source of competitive advantage for the company. The approach is inspired from the MEDIE model developed by [7]. Potential activities to outsource (or to back-source) are evaluated against three aspects: (1) current contribution of the activity to the competitive advantage, (2) potential contribution of the activity to the competitive advantage in the future, and (3) the sustainable performance of the activity. To perform the first two evaluations, we combine the value chain analysis and the Supply Chain Operations Reference model (SCOR). To evaluate sustainability performance, we first elect sustainability criteria and indicators based on the literature and GreenSCOR. Second, the Analytic Hierarchy Process (AHP) is used to assess the activities’ sustainability performance. Finally, the three evaluations are compared with each other and, based on the MEDIE model, different strategic decisions and options are proposed.

Keywords: make or buy, CSR, competitive advantage, value chain, SCOR, GreenSCOR, sustainability performance, AHP

INTRODUCTION Many activities historically integrated by companies have been progressively transferred to third parties. Outsourcing involves more and more complex functions of the value chain such as logistics [1]. Lastly, companies have also witnessed the advent of public requirements on product traceability (sourcing countries, carbon footprint, etc.) as well as the tightening of legislation in regard to environmental issues (GHG emissions, recycling policies, etc.). Corporate Social Responsibility (CSR) presents an attractive means for companies to address these requirements. From the corporate strategy perspective, CSR is considered as a differentiating factor for customers and other stakeholders [2]. According to [3], it is more and more difficult for companies relying on traditional differentiating means to remain competitive. Therefore, companies turn towards intangible factors such as reputation, legitimacy, corporate culture, innovation, etc. These differentiating sources would result from implementing CSR practices [3]. Therefore, value chain configuration decisions (e.g., outsourcing), which were traditionally based on economic

Integrating Corporate Social Responsibility …

87

criteria (e.g., cost, quality, and lead time), require, in the context of a differentiating strategy based on CSR practices, a new analysis that considers the contribution of the activities in creating CSR-based competitive advantage. In the literature, most studies attempting to integrate sustainability aspects in the supply chain configuration decisions tackle the problem from the logistics perspective solely. Typically, many mathematical models have been proposed to address the problem of “sustainable supply chain design” [4]. However, these studies do not take into account the company’s strategy. Moreover, a few studies addressed the decision of “make or buy”. Yet, this is a major strategic issue for companies. The motivation for outsourcing is that it is not always beneficial for organizations to make all activities in-house [5]. Thus, a company will always outsource a nonperforming or non-strategic activity, or internalize an external activity required for achieving its strategic goals [6]. Therefore, in a constantly changing environment, the company must adapt and continuously address the questions of “make, buy or ally” [1]. This chapter proposes a multi-disciplinary and multi-criteria approach for the problem of “make or buy” in the context where CSR is considered a source of competitive advantage. We also consider the options of backsourcing or sharing activities with other partners. To the best of our knowledge, this problem has not been tackled in the literature before. The proposed approach is inspired from the MEDIE model developed by [7]. Potential activities to outsource (or to back-source) are evaluated against three aspects: (1) the current contribution of the activity in creating a competitive advantage, (2) the potential contribution of the activity to competitive advantage in the future, and (3) the sustainable performance of the activity. To perform the two first evaluations, we combine the value chain analysis [8] and the Supply Chain Operations Reference model (SCOR) best practices and performance indicators. To evaluate the sustainable performance, we first elect sustainable criteria and indicators based on the literature. Second, based on the elected indicators, we use the Analytic Hierarchy Process (AHP) to assess the activities’ sustainability

88

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

performance. Finally, inspired by the MEDIE model, different strategic decisions and options are proposed. The remainder of the chapter is structured as follows. The next section presents a literature overview on the problem of make or buy and emphasizes models proposed in the context of sustainability operationalization. Section 3 briefly describes Porter’s value chain and the SCOR/GreenSCOR reference model. Section 4 presents the proposed approach. Finally, a conclusion and future research perspectives are presented in Section 5.

LITERATURE OVERVIEW The problem of make or buy consists in making the decision of outsourcing an activity (buy) or keeping the activity inside the company (make) [9, 10]. This decision might also concern an outsourced activity that the company might want to back-source [11]. In [12], it was stated that the make or buy decision consists of three distinct governance modes: Make, Buy, or Hybrid. The latter means that the activity is performed in partnership with another company such as a joint-venture. In [1], the hybrid governance mode is presented as the “ally” decision (i.e., sharing an activity with a partner). In this work, we consider all aforementioned decisions (make, buy, back-source, and ally). For the sake of simplicity in the reminder of the text, we refer to all these decisions as the problem of make or buy. This multidisciplinary problem has been studied from different perspectives such as economics, management, and logistics. A few works attempted to address the problem from a sustainability operations perspective. Our literature review is classified into two groups: Traditional models that do not consider sustainability aspects, and “Sustainability” models.

Integrating Corporate Social Responsibility …

89

Traditional Models The theory of transaction costs (e.g., [13, 14]) and the theory of resources (e.g., [15, 16]) are the most popular approaches. Within the first approach, a transaction is defined as product or service transfer between two technically separated entities. The organizational model (i.e., make or buy) to choose for organizing a transaction should minimize the sum of production and transaction (i.e., contract draft and negotiation, renegotiation, etc.). The second approach considers the make or buy decision from the perspective of the core of competencies. The competitive advantage depends, among others, on the resources that a company identifies, develops, deploys, and protects. These resources can be directly accessed by performing the activities in-house, or indirectly, by developing a partnership with other companies such as outsourcing. In [17], the author proposed a model that considers three factors: the technological process and its contribution to competitiveness, the maturity of this process compared to technologies available in the market, and the positioning of the technological process compared to competitors’ technologies. As an example, in the case where the technologies contribute to competitiveness and they are mature and well developed in the market, outsourcing would be the best option. In [18], the make or buy decision is assessed based on three criteria: the strategic importance of the activity (does it belong to the core of competencies?), the risk of dependency on partners, and the cost of performing the activity (in-house or outside the company). For instance, if an activity does not belong to the core of competencies but the risk of dependency is high, it should be performed in-house. In [7], the authors proposed a model called MEDIE (Modèle d’Evaluation de la Décision d’Internalisation/Externalisation in French, i.e., make or buy decision assessment model). This model is based on three criteria: the importance of the activity in creating the competitive advantage, the performance of the activity, and the potential of the activity to create competitive advantage in the future. The combination of the three criteria leads to one to multiple options. For instance, if the activity is not important and its performance is low but it has a potential to contribute to competitive advantage, the

90

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

company could cease the activity, outsource it, or perform it with a partner for whom the activity is important. The result of each one of the three evaluations is represented by a plus (+) (positive result) or a minus (-) (negative result) on the left and right side of a pastille, respectively. In [6], the author proposed a sequencing approach based on three analyses: operational feasibility of outsourcing (e.g., product handling, packaging, deterioration, etc.), strategic analysis, and cost analysis. In [19], a six-step outsourcing process is described: 1) analysis of the competencies, 2) evaluation and approval, 3) negotiation and contractualization, 4) project execution and transfer, 5) partnership management, and 6) partnership ending. At each step, the authors identify three variables; key activities, performance indicators, and outputs. Three main phases can be distinguished within the process: the strategic phase that addresses the questions “why?”, “what?”, and “who?” the transition phase that addresses the question “how?” and the operationalization phase that addresses the question of “how to manage the partnership” In [20], the authors used the SCOR model to map the activities of a large group in the chemical sector to identify which logistics activities should be outsourced. A three-step approach is proposed: 1) initial situation description and analysis to identify the activities to outsource (as is step), 2) future situation description and analysis, e.g., expectations from the future partner (to be step), and 3) outsourcing implementation and partnership management (go live step). In [21], a method for the re-configuration of a production system manufacturing personalized and high-end products in a global context was proposed. The proposed method is structured following four steps: product structure analysis to identify the core of competencies and activities to outsource, international cooperation analysis to identify the resources and communication means required for cooperating with the partners, analysis of the appropriate technologies in the different countries, and the final production and logistics system design. The authors separated the studied production system into three activity types: sourcing activities that should be located in countries offering a cost-based competitive advantage, activities manufacturing standard components that should be performed in-house and located in the country of origin, and assembly

Integrating Corporate Social Responsibility …

91

activities producing personalized end products, which should be located close to the market. Operational Research (OR)-based models have also been developed to tackle the problem of make or buy. In [22], the authors used queuing and Markov chain theories to model a logistics network composed of a contracting company and its multiple subcontractors. The authors assessed, in the short and long terms, the impact of different outsourcing strategies on the safety stock annual cost of the contracting company and customer satisfaction. The authors in [5] and [23] combined strategic and technical analyses to identify potential activities to outsource with an AHP-based algorithm to select potential sub-contractors, and different variants of mixed integer linear programming (MILP) models to design the optimal logistics network and identify the final activities to outsource, final subcontractors to select and the quantities of products to source from the subcontractors. In [24] and [25], the problem of make or buy was studied in the healthcare sector. The authors considered a sterilization service and analyzed three options: performing the activities in-house, outsourcing the service or mutualizing the service with other sterilization services. The authors used the value chain concept of Porter (more details are provided in Section 3) to identify potential sterilization activities to outsource. In [24], an MILP model was used to identify which activities to outsource and which partners to select. In [25], another MILP model was developed to determine the location of the sterilization service in the case of mutualization as well as its capacity and the optimal allocation of materials and human resources. In [26], a fuzzy multi-objective model that minimizes total cost of make and buy options, holding cost and rework cost in high-tech industry was proposed. The study in [27] proposes a three-stage optimization model that considers outsourcing and production decisions in the context of a global supply chain and China’s exportoriented tax policies. The decisions include determining which components should be imported and which quantities of products should be assigned to local and global markets.

92

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

Sustainability Models We identified a few models within this category ([28, 29, 30, 31, 32]). The approaches developed are OR-based and logistics-oriented. In [28], the author analyzed eight different reverse-logistics network configurations, each representing a combination of different logistics options (product collection, sorting and treatment). For instance, product collection can be performed by the company that collects only its own products or by another company/organization that collects the products of multiple companies while product sorting can be centralized (one point for product collection and sorting) or decentralized (within multiple collection points or in distinct points close to the collection points). AHP was used in [28] and [29] to select the best logistics network configuration. Two main criteria are considered; costs and business relationships, and each criterion encompasses multiple sub-criteria. In [28] an MILP model was developed to design the optimal logistics network including forward and reverse material flows. The demand and the volume of returned products are assumed stochastic. The author also used the MILP model to evaluate the cost of the logistics network configuration selected by using the AHP method. In [31] the Analytic Network Process (ANP) was used to assess four strategies of product collection in the solar energy industry: the company itself could collect the products, it could perform this operation with other partners, a logistics provider can collect the products, or the operation could be done by a governmental organization within a deposit-return system mechanism. The authors considered four criteria according to the BOCR model (Benefit, Opportunity, Cost, and Risk) [33] and identified 20 sub-criteria including socio-economic aspects (e.g., brand image, energy use, and recycling culture). In [30], the authors analyzed four transportation and warehousing options in a distribution network. They considered developing in-house both transportation and warehousing activities, developing in-house transportation and outsourcing warehousing, outsourcing transportation and developing warehousing inhouse, and finally, outsourcing both transportation and warehousing. They

Integrating Corporate Social Responsibility …

93

developed a multi-objective MILP model to design the optimal logistics network. The objective functions were to minimize total costs and to minimize different pollutant emissions (carbon, sulphur dioxide, pollutant particles, etc.) A multi-objective optimization model was developed in [32] to take into account the operational cost, capacity flexibility and outsourcing risk of a textile manufacturer. The authors suggest that reserving capacity flexibility with local suppliers and managing outsourcing risks related to international suppliers helps to achieve a sustainable outsourcing strategy. From the literature review, we can conclude the following: 





In most studies, the strategic analysis of the make or buy decision is not explicitly integrated with an analytic analysis within an integrated approach that selects potential activities to outsource (or to share with other partners) from the strategic perspective (e.g., analysis of the core of competencies or the competitive advantage) and analytically assesses the relevance of outsourcing the activities from a logistics perspective (e.g., costs). We identified only four works that proposed such an approach ([5, 6, 23, 24]). Works that proposed an integrated approach do not consider sustainability aspects in the strategic analysis or in the analytic evaluation of the make or buy decision. As stated before, CSR practices can contribute to creating a competitive advantage for companies that focus on intangible differentiating factors (reputation, legitimacy, corporate culture, etc.). Therefore, a new analysis of the make or buy decision is required in this context. Studies that considered sustainability aspects are very scarce. They focus on reverse logistics and distribution activities, and address only the analytic aspect of the problem. They lack a strategic analysis.

94

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

ACTIVITY MODELLING BY USING THE VALUE CHAIN AND SCOR In this section, we briefly describe the value chain concept of Porter [8] and the SCOR/GreenSCOR reference model. Then, we show the value of combining these two complementary tools for the problem of make or buy.

Value Chain Concept Developed by Michael Porter [8], the value chain is a strategic analysis tool that allows a company to identify its competitive advantage(s) and strengthening means to enhance it. According to Porter [34], a given company can gain a competitive advantage by performing its activities at lower costs (cost-based strategy) or better (differentiating strategy) than its competitors. From a strategic perspective, the value chain structures the company’s activities in a way that supports understanding cost behaviour and identifies existing and potential differentiating factors. The value chain comprises activities that create value and the margin (Figure 1). Activities creating value are of two types: primary activities and support activities. For instance, in the manufacturing sector, former activities concern raw material sourcing, the physical creation and sale of products, transportation up to the customers, and after-sale services. The latter support the primary activities, for example, by purchasing production means and providing technologies and human resources. Porter states that any company can be decomposed into five primary activities and four support activities as shown in Figure 1. Each activity can be decomposed into sub-activities. For instance, marketing and sales includes, among others, sales force management and advertising.

Integrating Corporate Social Responsibility …

95

Figure 1. Value chain [8].

SCOR/GreenSCOR Reference Model SCOR is a worldwide recognized framework for evaluating supply chain activities and performance [35]. It was developed by the Supply Chain Council (SCC) in 1996 (SCC merged with APICS in 2014 and the two now form APICS SCC). SCOR focuses on three “industry neutral” hierarchical levels as shown in Figure 2 (level four is specific to each industry and is not in SCOR scope). At the top level, six types of processes are identified: Source, Make, Deliver, Return, Plan, and Enable. The second level identifies the operations strategy. For example, the Make process can present the configurations Make-to-Stock, Make-to-Order, Engineer-to-Order, etc. The third level describes more in detail the activities comprised in the processes (process elements), e.g., “receive, enter and validate a customer order” [36]. Finally, a set of performance indicators (i.e., Metrics) and best practices are associated with each process and its elements. A practice is considered as “a unique way to configure a process or set of processes.” [36]. The uniqueness can be related to a technology or specific skills applied to the process, the automation of the process, the unique sequence of performing the process, etc. [36].

96

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

In Version 9.0 of SCOR, new best practices, performance indicators and activities related to environmental management have been introduced, leading to the creation of GreenSCOR. This new reference model, for instance, introduced the activity “Remove waste materials” within Make sub-processes. It also introduced 100 best practices and five performance indicators related to environmental management in the supply chain. As an example, “use alternative transportation modes consuming less fuel”, “reduce packaging use”, “reuse used packaging”, and “select transporters respecting environmental requirements” are environmental best practices related to Return process. The five performance indicators are quantity of CO2 emissions, quantity of pollutant particles emissions, quantity of liquid wastes, quantity of solid wastes, and % of recycled wastes.

Figure 2. Supply chain operations reference model [36].

Coupling the Value Chain with SCOR/GreenSCOR By comparing the value chain’s functional decomposition of activities with SCOR process-based decomposition, we can observe that SCOR processes can be easily represented within the value chain. As examples, Source process is equivalent to inbound logistics, Make is equivalent to

Integrating Corporate Social Responsibility …

97

operations, Deliver corresponds to outbound logistics, Return can be included in services, and Enable include Human Resources management.. To benefit from both approaches we propose to use the value chain to perform the strategic analysis of the company’s activities while including, where possible, SCOR sub-processes and activities. Our goal is twofold: 



The value chain is used for strategic analysis (e.g., identification of differentiating factors) which is not the main purpose of SCOR. This analysis is essential for make or buy decision. By integrating SCOR activities into the value chain, SCOR/GreenSCOR performance indicators and best practices (especially those related to CSR and sustainability performance) can easily be associated to the value chain activities.

This activity modelling perspective is used further (next section) in the approach proposed in this work.

PROPOSED APPROACH As mentioned earlier, our approach is inspired from the MEDIE model [7]. As in [7], we consider three evaluations of the activities. 1) The importance of the activity in creating the competitive advantage, 2) the performance of the activity, and 3) the potential of the activity to create a competitive advantage in the future. The results of these three evaluations are presented by (+) for a positive result and (-) for a negative result. In [7], the authors do not propose specific tools or techniques to perform these three evaluations. Moreover, the MEDIE model is not designed to cope with sustainability aspects. We recall that this work focuses on companies that implement CSR practices as strategic differentiating factors for creating a competitive advantage for customers and stakeholders [2, 3] (see Section 1).

98

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

Evaluation of the Activity’s Strategic Importance Table 1. Examples of uniqueness factors [34] based on CSR practices

Uniqueness factor categories

Examples of uniqueness factors based on SCR practice implementation (and related activities in the value chain)

Characteristics and performance of the products/services



Technologies used for the activity



Discretionary measures

 

 Work procedures





Linkages in the supply chain

Transportation modes consuming less fuel or electric transportation technologies (Inbound and Outbound logistics, Services, Technology development) Clean manufacturing technologies (Operations, Technology development) Work procedures ensuring a high security level for employees (Inbound and Outbound logistics, Operations) Environmental risk mitigation procedures ensuring a high security level for local communities, employees, and the environment (Inbound and Outbound logistics, Operations)

Linkage with suppliers Linkage between activities



Locally-based suppliers (Inbound logistics, Procurement)



Activities performed in the same or close sites (Operations)

Linkage with customers (distributors)



Responsible distributors (Outbound logistics, Services) Locally-based carriers (Outbound Logistics)

 Learning  Institutional factors

Fair trade products or locally-made products (Inbound logistics and Procurement) Recycled products (Operations, Procurement) Low carbon footprint products (Inbound and Outbound logistics, Procurement, Operations)





Well trained employees regarding environmental management (e.g., fuel use in transportation activities, security procedures) (Human Resources, Inbound and Outbound logistics, Operations) Employees trained in accident risk management (Human Resources, Operations) Social dialogue (e.g., with unions and local communities) and participatory planning with the stakeholders (Firm infrastructure)

According to Porter [34], in a differentiating-based strategy, a company differentiates itself from other companies when it offers something unique that surpasses simply offering a low price, and to which

Integrating Corporate Social Responsibility …

99

customers (and stakeholders) attach value. In a context where the customers and stakeholders have a demand for more sustainable and responsible operations, CSR practices can provide this unique competitive advantage. Therefore, this evaluation focuses on identifying CSR practices related to the activities that make these activities contribute to creating this unique competitive advantage. Inspired by Porter [34], we present in Table 1 examples of uniqueness factors based on CSR practices (inspired from the literature and GreenSCOR) that we link to the value chain activities. Figure 3 illustrates how CSR practices associated to “inbound logistics”, “operations”, “outbound logistics”, and “procurement” can be represented by using the value chain. These CSR practices are inspired from GreenSCOR (best practices) and the literature. When this evaluation is performed, it is assumed that such practices have already been implemented. Note that not all CSR practices lead to uniqueness and competitive advantage.

Evaluation of the Activity’s Potential to Create a Competitive Advantage Here, the activities (candidate for make or buy decision) are evaluated regarding their potential to contribute to the competitive advantage in the future. Similarly to previous evaluation, the value chain can be used to represent CSR practices that could be implemented in the future to analyze activities’ potential to contribute to creating a competitive advantage. To support this evaluation the SWOT (Strengths, Weaknesses, Opportunities, and Threats) tool could be used. A high performance R&D department (e.g., capable of rapidly developing new technologies or methods that significantly reduce energy consumption) is an example of an internal factor that positively impacts environmental impact reduction. The R&D department would be a strength for the company. Legislation regarding recycling or GHG emission reduction is an example of an external factor that could be seen as an opportunity to implement CSR practices and improve the activities’ sustainability performance.

100

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

Figure 3. Representation of activities’ differentiating factors (i.e., CSR practices) by using the value chain [8].

Evaluation of the Activity’s Sustainability Performance To evaluate the performance of the activities, we consider all three sustainability aspects: economic, environmental, and social. To this end, inspired from [40], we first elected 12 generic sustainability criteria; five within the economic dimension, four within the environmental dimension and three within the social one (Table 2). These criteria were also used in [4] and [41]. As an example, for each criterion, we specify an objective and give an example of a performance indicator related to that objective for operations activity of the value chain in Table 2. The 12 proposed criteria apply to all activities while the objectives and indicators might change depending on the activity considered. Based on the indicators specific to each activity, the (global) sustainability performance of the activity is then measured. Different methods can be used. As in [43], we use the AHP method. AHP addresses heterogeneity issues (e.g., measurement units of the indicators) and considers quantitative as well as qualitative indicators, among others. The software ExpertChoice

Integrating Corporate Social Responsibility …

101

is used in practice to implement AHP. The reader can refer to [45] for an example of the use of AHP and ExpertChoice in a real case study. Table 2. Sustainability performance criteria and examples of objectives and indicators related to operations activity Dimensions

Criteria 1. Financial performance 2. Responsiveness

Examples of objectives (Operations activity) Decrease production cost Decrease lead time

3. Reliability

Increase forecast reliability

4. Flexibility

Improve production flexibility

5. Quality

Improve products quality Decrease energy consumption during the manufacturing process Decrease pollutant emissions attributed to manufacturing activities Decrease GHG emissions attributed to manufacturing activities

Economy (Eco)

1. Resources use

2. Pollution Environment (Env)

Society (Soc)

3. GHG emissions

4. Hazardous materials 1. Health and security 2. Job creation and wealth

Eliminate hazardous wastes Preserve employees’ health and security Promote local employment

3. Work conditions

Offer good work conditions

Examples of indicators Production cost (SCOR) Production lead time (SCOR) Gap between production forecast and real production (i.e., forecast errors) (SCOR) Number of days required to achieve an increase of 20% of the normal production level (SCOR) % of rejected products [40] Quantity of energy consumed during the manufacturing process Quantity of pollutant particles released in the air [42] Quantity of CO2 emissions released during manufacturing (GreenSCOR) % of hazardous wastes eliminated [43] Accident frequency at work [44] % of locally-hired employees assigned to production activities [40] Level of salary maintenance in disease case [40]

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

102

The calculation procedure to evaluate the sustainability performance of candidate activities for make or buy decision is as follows: + − Let 𝐼𝑖𝑘𝑗𝑡 (𝐼𝑖𝑘𝑗𝑡 ) be the performance indicator i of activity k, within

sustainability dimension j, at period t, which improves the performance objective related to indicator i when its value increases (decreases, respectively). For the sake of simplicity and without loss of generality, we assume here that only one objective is associated to each sustainability criterion, and the objective is measured by only one indicator. Therefore, the couple “indicator-sustainability dimension” (i, j) belongs to the set {(1, Eco), (2, Eco), (3, Eco), (4, Eco), (5, Eco), (1, Env), (2, Env), (3, Env), (4, Env), (1, Soc), (2, Soc), (3, Soc)} (see Table 2) and k ranges from 1 to m where m is the number of candidate activities for make or buy decision. + + + − − − In addition 𝐼𝑖𝑘𝑗𝑡 ∈ [𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓 , 𝐼𝑖𝑘𝑗𝑡,𝑆𝑢𝑝 ] and 𝐼𝑖𝑘𝑗𝑡 ∈ [𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓 , 𝐼𝑖𝑘𝑗𝑡,𝑆𝑢𝑝 ]. + − Where 𝐼𝑖𝑘𝑗𝑡,𝑆𝑢𝑝 (𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓 ) is

the

target

of

the

+ indicator 𝐼𝑖𝑘𝑗𝑡

− + − (𝐼𝑖𝑘𝑗𝑡 respectively) and 𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓 (𝐼𝑖𝑘𝑗𝑡,𝑆𝑢𝑝 ) its lower (upper) bound. + − 𝐼𝑛𝑖𝑘𝑗𝑡 and 𝐼𝑛𝑖𝑘𝑗𝑡 are defined as the normalized indicators; that is the + − values of 𝐼𝑛𝑖𝑘𝑗𝑡 and 𝐼𝑛𝑖𝑘𝑗𝑡 range between 0 and 1.

𝐼𝑠𝑘𝑗𝑡 is the performance sub-index of activity k for sustainability dimension j, at period t. Its value ranges between 0 and 1. For instance, if j represents the environmental dimension (Env), then 𝐼𝑠𝑘𝑗𝑡 gives the environmental performance of activity k. 𝐼𝑔𝑘𝑡 is the global performance index of activity k, at period t. Its value ranges between 0 and 1. This index evaluates the sustainability performance of activity k. + Finally, we define 𝑤𝑖𝑗 as the weight of the normalized indicator 𝐼𝑛𝑖𝑘𝑗𝑡 − (𝐼𝑛𝑖𝑘𝑗𝑡 ) and 𝑊𝑗 the weight of sub-index 𝐼𝑠𝑘𝑗𝑡 .



+ − Step 1: for a given activity 𝑘, consider its indicators 𝐼𝑖𝑘𝑗𝑡 and 𝐼𝑖𝑘𝑗𝑡 .



+ − Step 2: normalize the indicators 𝐼𝑖𝑘𝑗𝑡 (𝐼𝑖𝑘𝑗𝑡 ) as follows:

Integrating Corporate Social Responsibility … 𝐼+

+ 𝐼𝑛𝑖𝑘𝑗𝑡 = 𝐼+ 𝑖𝑘𝑗𝑡

+ − 𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓

+ 𝑖𝑘𝑗𝑡,𝑆𝑢𝑝 − 𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓

− 𝐼𝑛𝑖𝑘𝑗𝑡 =1−



− − 𝐼𝑖𝑘𝑗𝑡 − 𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓 –

− 𝐼𝑖𝑘𝑗𝑡,𝑆𝑢𝑝 − 𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓

103 (1)

(2)

+ − Step 3: calculate the weight 𝑤𝑖𝑗 of each indicator 𝐼𝑛𝑖𝑘𝑗𝑡 (𝐼𝑛𝑖𝑘𝑗𝑡 )

(following AHP method).  For each sustainability dimension j ∈ {Eco, Env, Soc}, build the matrix 𝐴𝑗 = (𝑛*𝑛) (𝑛 ∈ {5,4,3}) where all possible pairs of indicators within dimension 𝑗 are considered to compare the indicators to each other within the pair. The indicators are presented as the elements of the rows and columns of the matrix 𝐴𝑗 and the results of the comparisons are the values of the cells (𝑖, 𝑖′) (𝑖. 𝑒. 𝑎𝑖𝑖 ) of the matrix (see Table 4 for an example). The comparisons are assumed to be performed by the managers, for instance, as described in [43]. A preference within AHP method is expressed by using a scale ranging from 1 to 9, where 1 represents equal importance and 9 indicates that the first element is nine times more important than the second element (see Table 7 for the description of all preferences). If an element 𝑖 in a given row of the matrix 𝐴𝑗 is “p times” more important than an element 𝑖′ in a given column, then 𝑖′ is “1/p time” more important than 𝑖. Therefore, for the cell (𝑖, 𝑖′), 𝑎𝑖𝑖′ = 1⁄𝑎𝑖′ 𝑖 and for the cell (𝑖,𝑖) 𝑎𝑖𝑖 = 1.  The weight 𝑤𝑖𝑗 of performance indicator 𝑖 is calculated as follows (AHP method):

𝑤𝑖𝑗 =

∑𝑖′

𝑎 ′ 𝑖𝑖 ∑𝑛 𝑎 ′ 𝑛𝑖

𝑛

(3)

Dyer [46] highlighted the issue of preference intransitivity within the AHP method. In fact, the pair comparisons can lead

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

104

to non-transitivity that cannot be eliminated. To address this issue, a consistency ratio of the judgements (CR), which gives an idea of how far is the distance between these judgements and the ideal model of perfect transitivity (calculated experimentally) is calculated: 𝐶𝑅 = 𝐶𝐼/𝑅𝐼

𝐼𝐶 is called the consistency index: 𝐼𝐶 =

(4) 𝜆𝑚𝑎𝑥 −𝑛 𝑛−1

where

𝜆𝑚𝑎𝑥 is the largest proper value of the judgement matrix 𝐴(𝑛 * 𝑛 ) and 𝑅𝐼 is a random consistency index, which is obtained for a matrix 𝐴′ of the same dimension as 𝐴 and for which 𝑎′𝑖′ 𝑖 = 1⁄𝑎′𝑖𝑖′ for all its elements 𝑖 and 𝑖 ′ . RC value should not be greater than 0.05 if the matrix dimension is (3*3), 0.08 if its dimension is (4*4) and 0.1 if its dimension is equal or greater than (5*5) [45], otherwise, the preferences and weights should be re-evaluated. 

Step 4: calculate sub-index 𝐼𝑠𝑘𝑗𝑡 for each sustainability dimension (i.e., economic, environmental and social performances). + − 𝐼𝑠𝑘𝑗𝑡 = ∑𝑖 𝑤𝑖𝑗 ∗ 𝐼𝑛𝑘𝑖𝑗𝑡 + ∑𝑖 𝑤𝑖𝑗 ∗ 𝐼𝑛𝑘𝑖𝑗𝑡

(5)



Step 5: calculate the weights 𝑊𝑗 of sub-indices 𝐼𝑠𝑘𝑗𝑡 (same procedure as in step 3).



Step 6: calculate the global index 𝐼𝑔𝑘𝑡 (i.e., sustainability performance of the activity)

𝐼𝑔𝑘𝑡 = ∑𝑗 𝑊𝑗 ∗ 𝐼𝑠𝑘𝑗𝑡

(6)

Integrating Corporate Social Responsibility …

105

Illustration of the Calculation Procedure on a Hypothetical Example Consider a manufacturing activity which is a candidate for make or buy decision, and for which the values of its economic, environmental and social indicators are collected for the previous year. To illustrate, we selected these indicators from Table 2. The collected values as well as the results of the sustainability performance evaluation of the activity following the procedure described above are presented in Table 3. Tables 4 to 6 illustrate the judgement matrices obtained for the economic, environmental, and social indicators, respectively. As mentioned earlier, these evaluations are assumed to be performed by the managers. A given value in the tables describes the preference of an indicator in the raw (criterion) over another indicator in the column. The description of the preferences is presented in Table 8. All values of the three judgement matrices have been normalized following formula (3) to obtain the weights of the indicators. The final results are presented in Table 3 (indicator weights). Moreover, the consistency ratios have been calculated following formula (4) for the judgement matrices of the economic and environmental indicators. These are respectively 0.0032 and 0.0079 (≤ 0.05). These results show that the judgements are consistent. Since only two indicators were considered for the social performance, calculating the consistency ratio is not required. Table 8 shows the judgement matrix for the three sustainability dimensions. Again, the preferences were normalized following formula (3) to obtain the weights of the three sustainability dimensions (see Table 3), and the consistency ratio was calculated as well (0.0079 ≤ 0.05). From Table 3, we can see that the value of the environmental subindex is the highest, which means that the evaluated activity has a higher environmental performance compared to its economic and social performances. To evaluate if the activity presents a sufficient sustainability performance, the global index is analyzed. Its value should be compared to the performance target set for the activity by the managers. For instance,

Table 3. Results of the sustainability performance evaluation of a manufacturing activity Indicators 𝑰−𝒊𝒌𝒋𝒕

Economy Production cost Production lead time Number of days required to achieve an increase of 20% of the normal production level Environment Quantity of energy consumed Quantity of CO2 emissions released % of hazardous wastes eliminated Society Accident frequency at work

Indicator** 𝑰+𝒊𝒌𝒋𝒕

Values

[𝑰−𝒊𝒌𝒋𝒕,𝑰𝒏𝒇 ; 𝑰−𝒊𝒌𝒋𝒕,𝑺𝒖𝒑 ]

Normalized indicator 𝑰𝒏−𝒊𝒌𝒋𝒕

Indicator weights 𝒘𝒊𝒋

100 $/p.u* 14 days/p.u 5 days

[80; 130] [7; 20] [2; 7]

0.600 0.460 0.400

0.084 0.472 0.444

0.024 toe/p.u 0.145 t/p.u

[0.010; 0.030] [0.115; 0.180]

0.300 0.540

0.082 0.343

4 kg/p.u

[1; 10]

0.670

0.575

1.7/million of worked hours

[1; 2.2]

0.420

0.857

Value

+ + [𝐼𝑖𝑘𝑗𝑡,𝐼𝑛𝑓 ; 𝐼𝑖𝑘𝑗𝑡,𝑆𝑢𝑝 ]

Normalized + indicator𝐼𝑛𝑖𝑘𝑗𝑡

Indicator weights 𝑤𝑖𝑗

Subindices 𝑰𝒔𝒌𝒋𝒕

Sub-index weights 𝑾𝒋

0.445

0.633

0.595

0.175

0.410

0.192

% of locally-hired employees 47% [40; 60] 0.350 0.143 * p.u: production unit (kg, m3, m2, etc.), toe: tonne oil equivalent, t: ton. ** There is only one indicator (social dimension which improves the social performance of the activity when its value increases).

Global index (sustainability performance) 𝑰𝒈𝒌𝒕

0.465

Integrating Corporate Social Responsibility …

107

Table 4. Judgement matrix for the economic indicators

Production cost Production lead time Number of days required to achieve an increase of 20% of the normal production level

Production cost

Production lead time

1 6 5

1/6 1 1

Number of days required to achieve an increase of 20% of the normal production level 1/5 1 1

Table 5. Judgement matrix for the environmental indicators

Quantity of energy consumed Quantity of CO2 emissions released % of hazardous wastes eliminated

Quantity of energy consumed 1

Quantity of CO2 emissions released 1/5

% of hazard wastes eliminated 1/6

5

1

1/2

6

2

1

Table 6. Judgement matrix for the social indicators

Accident frequency at work % of locally-hired employees

Accident frequency at work 1 6

% of locally-hired employees 1/6 1

Table 7. Preferences and description Prefrence 1 3 5 7 9 2, 4, 6, 8

Description Identical preference Moderate preference of a criterion over another Strong preference of a criterion over another Very strong preference of a criterion over another Extreme preference of a criterion over another Intermediate values

108

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle Table 8. Judgement matrix for sustainability dimensions Economy 1 1/4 1/3

Economy Environment Society

Environment 4 1 1

Society 3 1 1

Analysis of the Three Evaluations and Decision-Making Table 9. Analysis grid for the decision of make or buy, adapted from [7] Activity evaluations Importance Performance Potentiel -*

Possible decisions Outsource

-

-

+

Outsource Share

+

-

-

Outsource Share Make in-house

-

+

+

Share Make in-house

-

+

-

Make in-house

+

-

+

Make in-house

+ +

+

Possible options Outsource the activity if it is essential Cease the activity Outsource the activity Perform the activity with a partner for whom the activity is important Outsource the activity Reengineer the activity/decompose/ re-assign the activity to another service, location, etc. or to a partner Focus on improving the activity performance Perform the activity with a partner for whom the activity is important Create a subsidiary Perform the activity for other business units of the company Cease the activity Focus on improving the activity performance Maintain and monitor the activity Pay strategic and constant attention to the activity

Make in-house Make inhouse/backsource *A plus (+) means that the activity achieved an acceptable result while a minus (-) means that the result is insufficient. + +

Integrating Corporate Social Responsibility …

109

if we assume in our example that a value below 0.5 is the minimum performance level, then we can conclude that the performance of the activity evaluated is insufficient since the global index obtained is 0.465. The results of the three evaluations are compared in this phase, and a strategic decision and an option are identified. The options range from ceasing to back-sourcing an activity. All possible combinations of the three evaluations (+ and – represent the results of the evaluations) and the resulting strategic decisions and options are summarized in the analysis grid presented in Table 9. This grid is adapted from the MEDIE model [7].

CONCLUSION Traditionally, governmental and public requirements for more sustainable businesses and operations were seen as constraints by companies. However, in a more and more competitive environment where it is difficult to differentiate from the competitors by traditional means, implementing CSR practices is considered as an opportunity to create competitive advantage. In this context, the decision of make or buy becomes more complex to evaluate. In the literature, traditional models as well as models proposed to cope with sustainability are not appropriate for dealing with this problem. The approach proposed in this work considers both the strategic importance of the activities (current and potential competitive advantage based on the implementation of CSR practices) and their sustainability performance (economic, environmental and social aspects). We also include in the study the possibilities of back-sourcing an activity or sharing it with a business partner. Given that the problem of make or buy is multi-disciplinary, we combined different management tools (value chain, SCOR, etc.) and engineering techniques (e.g., AHP). The analysis proposed at the final phase of the approach leads to multiple strategic decisions and options that could be further refined. To select the final option, an optimization or a simulation model could be developed to consider for example logistics costs and GHG emissions. Our proposed approach could be seen as a strategic filter that can be performed prior to

110

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

this logistics analysis. Finally, applying the approach on a real case study would be very interesting to show the practical value of the approach and identify improvement opportunities.

ACKNOWLEDGMENT Funding for this project was provided by La Région Rhône-Alpes.

REFERENCES [1]

[2]

[3]

[4]

Quélin, Bertrand. 2003. “Externalisation stratégique et partenariat : de la firme patrimoniale à la firme contractuelle ?” [Strategic outsourcing and partnership: from the patrimonial firm to the contractual firm] Revue Française de Gestion 143(2):13-26. https://www.cairn.info/revue-francaise-de-gestion-2003-2-page13.htm. Hoffmann, Gérard and Jean-Yves Saulquin. 2009. “Quand la RSE revisite la chaîne de valeur.” [When the CSR revisits the value chain] Management & Avenir 28(8):37-55. https://www.cairn.info/revuemanagement-et-avenir-2009-8-page-37.htm. Li, Ning and Anne Toppinen. 2011. “Corporate responsibility and sustainable competitive advantage in forest-based industry: Complementary or conflicting goals?” Forest Policy and Economics 13:113-123. doi:10.1016/j.forpol.2010.06.002. Boukherroub, Tasseda, Angel Ruiz, Julien Fondrevelle, Alain Guinet and Mustapha Ouhimmou. 2016. “Intégration de la performance durable dans les modèles de conception et de planification des chaînes logistiques: un état de l’art” [Integrating the sustainable performance into supply chain design and planning models: A state of the art]. Journal Européen des Systèmes Automatisés 49(3):237270. doi: 10.3166/JESA.49.237-270.

Integrating Corporate Social Responsibility … [5]

[6]

[7]

[8] [9] [10]

[11]

[12]

[13] [14] [15]

111

Bouchriha, Hanen. 2012. “Faire ou faire-faire dans la conception d’une chaine logistique : un outil d’aide à la décision.” [Make or buy in supply chain design:a decision support tool.] PhD diss., Institut National Polytechnique de Grenoble. Lakhal, Salem. 1998. “Vers une approche formelle d’aide à la décision dans l’entreprise réseau” [Towards a formal decision support approach for the network enterprise] PhD diss., Université Laval. Poulin, Diane, Benoit Montreuil and Stéphane Gauvin. 1994. “L’entreprise réseau, bâtir aujourd’hui l’organisation de demain” [The network enterprise: build today the organization of tomorrow]. Montreal: Publi-Relais. Porter, Michael. 1985. The Competitive advantage of nations. New York: Free Press. Barreyre, Pierre Yves. 1968. L’impartition, politique pour une entreprise compétitive. Paris: Hachette. Foss, Nicolai J. 1996. “Capabilities and the Theory of the Firm.” Revue d’Economie Industrielle 77:7-27. doi: https://doi.org/10.3406/ rei.1996.1633. Fréry, Frédéric and Law-Kheng Florence. 2007. “La réinternalisation, chaînon manquant des théories de la firme.” Revue Française de Gestion 177(8): 163-179. doi:10.3166/RFG.177.163179. Jahns Christopher, Evi Hartmann and Lydia Bals. 2006. “Offshoring: Dimensions and diffusion of a new business concept.” Journal of Purchasing and Supply Management 12(4): 218-231. doi:10.1016/j. pursup.2006.10.001. Coase, Ronald H. 1937. “The nature of the firm.” Economica 4:386405. doi:10.1111/j.1468-0335.1937.tb00002.x. Williamson, Oliver E. 1985. The Economic Institutions of Capitalism. New York: Free Press. Wernerfeldt, Birger. 1989. “From critical resources to corporate strategy.” Journal of General Management 14(3): 4-12. doi:10.1177/ 030630708901400301.

112

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

[16] Barney, Jay. 1991. “Firm resources and sustained competitive advantage.” Journal of Management 17:99-120. doi:10.1177/ 014920639101700108. [17] Welch, James M. and P. Ranganath Nayak. 1992. “Strategic Sourcing: A progressive approach to make or buy decision.” The Executive 6(1):23-31. https://www.jstor.org/stable/4165048. [18] Brandes, Henrik. 1994. “Strategic changes in purchasing, two main tracks.” European Journal of Purchasing and Supply Chain 1(2):7787. doi:10.1016/0969-7012(94)90018-3. [19] Momme, Jasper and Hans-Henrik. Hvolby. 2002. “An outsourcing framework: action research in the heavy industry sector.” European Journal of Purchasing & Supply Management 8:185-196. doi:10. 1016/S0969-7012(02)00003-5. [20] Wattky Crestan, Andrea. 2006. “Modélisation et Réorganisation des Processus dans le Cadre de l’Externalisation” [Process modelling and reorganization in outsourcing context.] PhD diss., Université Lumière Lyon 2. [21] Wagner, Carsten and Peter Nyhuis. 2009. “A systematic approach to analysis and design of global production network.” Production Engineering Research Development 3:295-303. doi:10.1007/s11740009-0167-z. [22] Layek, Abdel-Malek, Tarathorn Kullpattaranirun and Suebsak Nanthavanij. 2005. “A framework for comparing outsourcing strategies in multi-layered supply chains.” International Journal of Production Economics 97:318-328. doi:10.1016/j.ijpe.2004.09.001. [23] Bouchriha, Hanen. 2009. “Aide à la décision pour la conception et le pilotage des systèmes de production de biens et de services” [Decision-making support for the design and planning of goods and services production systems] Accreditation to direct research, Ecole Nationale d’Ingénieur de Tunis. [24] Tlahig, Houda, Hanen Bouchriha, Aida Jebali and Pierre Ladet, 2007. “A mathemathical model for the “make” or “buy” decision in the hospital sterilization.” Paper presented at the 33rd International

Integrating Corporate Social Responsibility …

[25]

[26]

[27]

[28] [29]

[30]

[31]

[32]

113

Conference on Operational Research Applied to Health Service, Saint-Etienne, France, July 15-20. Tlahig, Houda, Hanen Bouchriha, Aida Jebali and Pierre Ladet. 2008. “Centralized Versus Distributed Sterilization Service: A Location-Allocation Decision Model.” Paper presented at the International Conference on Information Systems, Logistics and Supply Chain, Madison, Wisconsin, May 28-30. Wang, Earl-Juei, Tai-Sheng Su, Deng-Maw Tsai and Chun-Yu Lin. 2013. “Fuzzy multiple-goal programming for analysing outsourcing cost-effectiveness in hi-tech manufacturing.” International Journal of Production Research 51(13):3920-3944. doi:10.1080/ 00207543.2012.757665. Zhen, Lu. 2014. “A Three-Stage Optimization Model for Production and Outsourcing under China’s Export-Oriented Tax Policies.” Transportation Research Part E 69:1-20. doi:10.1016/j.tre. 2014.05.014. Barker, Theresa J. 2010. “Reverse logistics: A multicriteria decision model with uncertainty.” PhD diss., University of Washington. Barker, Theresa J. and Zelda B. Zabinsky. 2011. “A multicriteria decision making model for reverse logistics using an analytical hierarchy process.” Omega 39: 558-573. doi:10.1016/j.omega. 2010.12.002. Mallidis, Ioannis, Rommert Dekker and Dimitrios Vlachos. 2012. “The impact of greening on supply chain design and cost: a case for a developing region.” Journal of Transport Geography 22:118-128. doi: 10.1016/j.jtrangeo.2011.12.007. Shiue, Yih-Chearng and Chun-Yueh Lin. 2012. “Applying analytic network process to evaluate the optimal recycling strategy in upstream of solar energy industry.” Energy and Buildings 54:266 277. doi: 10.1016/j.enbuild.2012.07.032. Sardar Shaheen, Young Hae Lee and Muhammad Saad Memon. 2016. “A Sustainable Outsourcing Strategy Regarding Cost, Capacity Flexibility, and Risk in a Textile Supply Chain.” Sustainability 8(3):234. doi:10.3390/su8030234.

114

Tasseda Boukherroub, Alain Guinet and Julien Fondrevelle

[33] Saaty, Thomas L., Müjgan Sagir Özdemir. 2003. “Negative priorities in the analytic hierarchy process.” Mathematical and Computer Modelling 37:1063-1075. doi: 10.1016/S0895-7177(03)00118-3. [34] Porter, Michael. 1999. L’avantage concurrentiel, Comment devancer ses concurrents et maintenir son avance [The competitive advantage, How to get ahead of your competitors and maintain your lead]. Paris: DUNOD. [35] APICS Supply Chain Council http://www.apics.org/about/overview/ about-apics-scc. Last accessed on May 2019. [36] SCC, Supply Chain Operations Reference Model, Supply Chain Council, 2012. [37] Barbiroli, Giancarlo and Andrea Raggi. 2003. “A method for evaluating the overall technical and economic performance of environmental innovations in production cycles.” Journal of Cleaner Production 11(4):365-374. doi: 10.1016/S0959-6526(02)00058-6. [38] Carter, Craig R. and Marianne M. Jennings. 2002. “Logistics social responsibility: An integrative framework.” Journal of Business Logistics 23(1):145-178. doi: 10.1002/j.2158-1592.2002.tb00020.x. [39] Corsi, Thomas and Philip Fanara. 1988. “Driver management policies and motor carrier safety.” Logistics and Transport Review 24(2):153-163. [40] Baumann, Émilie. 2011. “Modèles d’évaluation des performances économique, environnementale et sociale dans les chaînes logistiques.”[Assessment models of the economic, environmental and social performances in supply chains] PhD diss., Institut National des Sciences Appliquées de Lyon. [41] Boukherroub, Tasseda, Angel Ruiz, Alain Guinet, Julien Fondrevelle. 2015. “An integrated approach for sustainable supply chain planning.” Computers & Operations Research, 54:180-194. doi:10.1016/j.cor.2014.09.002. [42] Ferreti, Ivan, Simone Zanoni, Lucio Zavanella and A. Diana. 2007 “Greening the aluminium supply chain.” International Journal of Production Economics 108:236-245. doi:10.1016/j.ijpe.2006.12.037.

Integrating Corporate Social Responsibility …

115

[43] Krajnc, Damjan and Peter Klavic. 2005. “A model for integrated assessment of sustainable development.” Resources, Conservation and Recycling 43(2):189-208. doi:10.1016/j.resconrec.2004.06.002 [44] Singh, Rajesh Kumar, H. Ramalinga Murty, Shivendu Kumar Gupta and Anil Kumar Dikshit. 2007. “Development of composite sustainability performance index for steel industry.” Ecological Indicators 7(3):565-588. doi: 10.1016/j.ecolind.2006.06.004. [45] Boukherroub, Tasseda, Luc L. LeBel and Angel Ruiz. 2017. “A framework for sustainable forest resource allocation: A Canadian case study.” Omega 66(B):224-235. doi:10.1016/j.omega. 2015.10.011 [46] Dyer, James S. 1990. “Remarks on the Analytic Hierarchy Process.” Management Science (36):249-258. doi:10.1287/mnsc.36.3.249.

In: Sustainable City Logistics Planning ISBN: 978-1-53616-561-6 Editor: Anjali Awasthi © 2020 Nova Science Publishers, Inc.

Chapter 5

WHICH DATA TO COLLECT AND HOW, SO AS TO UNDERSTAND URBAN FREIGHT DISTRIBUTION? SETTING THE FRAMEWORK Katrien De Langhe, PhD, Roel Gevaers, PhD and Thierry Vanelslander*, PhD Department of Transport and Regional Economics University of Antwerp Antwerp, Belgium

ABSTRACT In order to deal with issues such as congestion and air pollution, authorities start introducing progressive measures to improve the liveability and sustainability in their urban area. However, policy makers often introduce these measures without analysing the context in which they are applied. By collecting data on a regular basis, authorities can make better decisions about their policy with respect to urban freight distribution. This study aims to contribute to this growing area of research by developing a framework of urban freight distribution *

Corresponding Author’s Email: [email protected].

118

Katrien De Langhe, Roel Gevaers and Thierry Vanelslander indicators and data collection methods. The framework demonstrates that parts of urban freight distribution are characterised by a set of indicators and cannot be captured by measuring just one indicator. This is a very important take-away for decision makers and other urban freight stakeholders. The framework also offers the possibility to see which indicators can be measured once a certain data collection method has been chosen. This enables decision makers to collect data in a focused way, leading to time- and financial benefits.

Keywords: urban freight distribution, data collection, indicators

INTRODUCTION Urban freight distribution is affected by multiple trends. The world population is growing, also in cities. In 2018, 55.4% of all people in the world were living in urban areas, whereas this is expected to increase to 66% by 2050 (United Nations, 2018). The increasing population leads to more person mobility, but also to more freight transport (European Commission, 2018). One result of the growing person mobility and freight transport is that urban areas face more economic and environmental issues, such as congestion and air pollution. In order to deal with these issues, authorities start introducing progressive measures to improve the liveability and sustainability in their urban area. Examples are the introduction of time windows and low emission zones. However, policy makers often introduce these measures without analysing the context in which they are applied (Stathopoulos, Valeri, & Marcucci, 2012). By collecting data on a regular basis, authorities can make better decisions about their policy with respect to urban freight distribution (Muñuzuri, Cortés, Onieva, & Guadix, 2018). The existing literature on this topic reveals that there is a lack of publicly available data on urban freight distribution (Julian Allen, Browne, & Cherrett, 2012; Campagna, Stathacopoulos, Persia, & Xenou, 2017; Cherrett et al. 2012; Dablanc, 2009). Subsequently, the literature demonstrates that freight transport is often neglected in existing models and surveys. If freight transport is part of the models and surveys, this is

Which Data to Collect and How …

119

often at an aggregated level and not at an urban level. If data are collected, the methodologies used for different data collection initiatives are often not comparable (Ambrosini & Routhier, 2004; Crainic, Ricciardi, & Storchi, 2004). Several authors have tackled the problem of urban freight data collection. Binnenbruck (2006) provides some examples of data collection methods and indicators for Germany. Routhier & Patier (2006), Browne et al. (2007) and Patier & Routhier (2009) offer some general data collection indicators and methods. Ban et al. (2010) as well as Holguin-Veras & Jaller (2012) provide a data collection framework. However, no specific urban freight indicators are given in these studies. Gonzalez-Feliu et al. (2013) show collection methods and a typology of indicators, related to shipment, pickup or delivery operation, vehicles and external elements. Routhier (2013) provides categories of collection methods, without describing indicators. Allen et al. (2014) make a comparison of urban freight data collected between countries. However, in all these works, no direct link is made between the urban freight indicators and the collection methods. Only a few authors have made the link between urban freight collection methods and indicators. Patier & Routhier (2008) propose some indicators depending on the objective of the data collection and mention the appropriate collection method. Allen et al. (2012) offer some survey techniques and 11 basic indicators. Pluvinet et al. (2012) match 13 indicators and five collection methods. However, in these papers, only a few broad indicators are given and no examples of data collection efforts measuring these indicators are provided. Moreover, Allen et al. (2014) state that stakeholders often do not know whether the indicators they use are commonly potential useful indicators. This study aims to contribute to this growing area of research by developing a framework of urban freight distribution indicators and data collection methods. The used method is based on a three-stage analysis. Firstly, an overview and analysis of literature on data availability is given. Subsequently, the literature is applied to a framework for urban freight profiles. Ultimately, a few case studies illustrate the theory. The remainder of this chapter is structured as follows. Section 2 provides the main

120

Katrien De Langhe, Roel Gevaers and Thierry Vanelslander

observations from the existing literature. Section 3 offers urban freight indicators, whereas Section 4 gives an overview of the data collection methods. In Section 5, a case study for Antwerp illustrates the use of the indicators and methods. In Section 6, some conclusions are drawn.

LITERATURE OVERVIEW The first author who published an extensive book on urban freight is Ogden (1992). Since the 21st century, the data availability of urban freight distribution has been equal, or has even slightly improved. This is the result of new national freight surveys in some countries, or data collection efforts at local urban level. One of the most important data collection initiatives in Europe took place in France around 1997. In 2011-2012, a new survey round was set up for the city of Paris (Toilier et al. 2016). This initiative was executed thanks to financing by the government (Browne et al. 2007). Browne et al. (2007), Patier & Routhier (2008), Ban et al. (2010), Holguin-Veras & Jaller (2012) and Allen et al. (2014) list the main gaps in available data. Concerning urban freight distribution, these main gaps are data concerning empty flows, activities of lorries