Multimodal Transport Systems

The use and management of multimodal transport systems, including car-pooling and goods transportation, have become extremely complex, due to their large size (sometimes several thousand variables), the nature of their dynamic relationships as well as the many constraints to which they are subjected. The managers of these systems must ensure that the system works as efficiently as possible by managing the various causes of malfunction of the transport system (vehicle breakdowns, road obstructions, accidents, etc.). The detection and resolution of conflicts, which are particularly complex and must be dealt with in real time, are currently processed manually by operators. However, the experience and abilities of these operators are no longer sufficient when faced with the complexity of the problems to be solved. It is thus necessary to provide them with an interactive tool to help with the management of disturbances, enabling them to identify the different disturbances, to characterize and prioritize these disturbances, to process them by taking into account their specifics and to evaluate the impact of the decisions in real time. Each chapter of this book can be broken down into an approach for solving a transport problem in 3 stages, i.e. modeling the problem, creating optimization algorithms and validating the solutions. The management of a transport system calls for knowledge of a variety of theories (problem modeling tools, multi-objective problem classification, optimization algorithms, etc.). The different constraints increase its complexity drastically and thus require a model that represents as far as possible all the components of a problem in order to better identify it and propose corresponding solutions. These solutions are then evaluated according to the criteria of the transport providers as well as those of the city transport authorities. This book consists of a state of the art on innovative transport systems as well as the possibility of coordinating with the current public transport system and the authors clearly illustrate this coordination within the framework of an intelligent transport system.

Table of Contents:
Preface xi Slim HAMMADI and Mekki KSOURI Chapter 1. Dynamic Car-pooling 1 Slim HAMMADI and Nawel ZANGAR 1.1. Introduction 1 1.2. State of the art 2 1.3. Complexity of the optimized dynamic car-pooling problem: comparison and similarities with other existing systems 8 1.3.1. Graphical modeling for the implementation of a distributed physical architecture 9 1.3.2. Collection of requests for car-pooling and data modeling 11 1.3.3. Matrix structure to collect information on requests 14 1.3.4. Matrix representation for modeling car-pooling offers 17 1.3.5. Modeling constraints of vehicles’ allocation to users 21 1.3.6. Geographical network subdivision served and implementation of a physical distributed dynamic architecture 26 1.4. ODCCA: an optimized dynamic car-pooling platform based on communicating agents 33 1.4.1. Multi-agent concept for a distributed car-pooling system 33 1.5. Formal modeling: for an optimized and efficient allocation method 40 1.5.1. D3A: Dijkstra Dynamic Distributed Algorithm 40 1.5.2. ODAVe: Optimized Distributed Allocation of Vehicle to users 44 1.6. Implementation and deployment of a dynamic car-pooling service 47 1.6.1. Deployment of ODCCA: choosing a hybrid architecture 49 1.6.2. Layered architecture 51 1.6.3. Testing and implementation scenario 55 1.7. Conclusion 65 1.8. Bibliography 66 Chapter 2. Simulation of Urban Transport Systems 71 Christian TAHON, Thérèse BONTE and Alain GIBAUD 2.1. Introduction 71 2.2. Context 72 2.3. Simulation of urban transport systems 75 2.3.1. Non-guided transport systems 76 2.3.2. Guided transport systems 77 2.4. The types of modeling 80 2.4.1. Nature of the models 80 2.4.2. Macrosimulation, mesoscopic simulation, micro simulation 81 2.5. Modeling approaches 83 2.6. Fields of application 83 2.7. Software tools 86 2.8. Simulation of the Valenciennes transport network with QUEST software 87 2.8.1. Problem 87 2.8.2. Network operation in normal mode 87 2.8.3. Disturbed mode network function 90 2.9. The QUEST software 92 2.9.1. Presentation 92 2.9.2. Modeling 92 2.10. Network modeling in normal mode 94 2.10.1. Topology of traffic networks 94 2.10.2. Bus lines 95 2.10.3. Vehicles 96 2.10.4. Modeling 96 2.10.5. Stops 97 2.10.6. Passengers 101 2.10.7. The flow of connecting passengers 102 2.11. Network modeling in degraded mode 103 2.11.1. Disturbances 103 2.11.2. Regulatory procedures 105 2.12. Simulation results 107 2.13. Conclusion/perspectives 107 2.14. Self-organization of traffic – the FORESEE simulator 108 2.14.1. General problem 108 2.14.2. FORESEE simulator 113 2.14.3. Results 117 2.15. Conclusion – perspectives 124 2.15.1. Sustainability of the information 125 2.15.2. Information aggregation algorithms 125 2.15.3. Cooperation efficiency 125 2.15.4. Deployment of the proposed approach 126 2.16. Bibliography 127 Chapter 3.Real-time Fleet Management: Typology and Methods 139 Frédéric SEMET and Gilles GONCALVES 3.1. Introduction 139 3.2. General context of RTFMP 140 3.2.1. RTFMP characteristics 140 3.2.2. Application field of RTFMPs 142 3.3. Simulation platform for real-time fleet management 144 3.3.1. Dynamic management of vehicle routing 144 3.3.2. Routing management under time window constraints 146 3.3.3. General architecture of the simulation platform 147 3.3.4. Consideration of uncertainties on requests 151 3.3.5. Consideration of information linked to traffic 156 3.4. Real-time fleet management: a case study 162 3.4.1. General architecture of the optimization engine 163 3.4.2. Itinerary calculation and length estimation 164 3.4.3. The static route planning problem 165 3.4.4. Route planning and modification of the transport plan 166 3.5. Conclusion 168 3.6. Bibliography 168 Chapter 4. Solving the Problem of Dynamic Routes by Particle Swarm  173 Mostefa Redouane KHOUAHJIA, Laetitia JOURDAN and El Ghazali TALBI 4.1. Introduction 173 4.2. Vehicle routing problems 174 4.2.1. The static vehicle routing problem 174 4.2.2. The dynamic vehicle routing problem (DVRP) 176 4.2.3. Importance of dynamic routing problems 178 4.3. Resolution scheme of the dynamic vehicle routing problem 179 4.3.1. Event planner 179 4.3.2. Particle swarm optimization 181 4.4. Adaptation of the PSO metaheuristic for the dynamic vehicle routing problem 184 4.4.1. Representation of particles 184 4.4.2. Velocity and movement of particles 185 4.4.3. The APSO algorithm (Adaptive Particle Swarm Optimization) 187 4.4.4. Adaptive memory mechanism 188 4.5. Experimental results 189 4.5.1. Datasets 189 4.5.2. Experiments and analysis 190 4.5.3. Measure of dynamicity 192 4.6. Conclusion 196 4.7. Bibliography 196 Chapter 5. Optimization of Traffic at a Railway Junction: Scheduling Approaches Based on Timed Petri Nets  199 Thomas BOURDEAUD’HUY and Benoît TROUILLET 5.1. Introduction 199 5.2. Scheduling in a railway junction 201 5.2.1. Classical scheduling 201 5.2.2. Flexible system scheduling 202 5.2.3. Dual Gantt diagram 203 5.2.4. The railway junction saturation problem 204 5.3. Petri nets for scheduling 206 5.3.1. Place/Transition Petri net 206 5.3.2. T-timed Petri nets 209 5.3.3. Controlled executions 211 5.3.4. Reachability problems in TPNs 212 5.3.5. Modeling of a railway junction with Petri nets 213 5.3.6. Approaches to solving the timed reachability problem 214 5.4. Incremental model for TPNs 216 5.4.1. Formulation operators “+” and “s” 220 5.4.2. Integer Mathematical Models 223 5.4.3. Numerical experiments 225 5.4.4. Study of the illustrative example of Figure 5.5 227 5.4.5. Conclusion and future work 228 5.5. A (max,+) approach to scheduling 229 5.5.1. Introduction and production hypotheses 230 5.5.2. Construction of a simple event graph associated with the initial model 233 5.5.3. Resolution of resource sharing 236 5.5.4. Application 242 5.5.5. Overview 246 5.6. Conclusion 247 5.7. Bibliography 248 List of Authors 253 Index 255

About the Author :
Slim Hammadi is Full Professor at the Ecole Centrale de Lille in France, and Director of the LAGIS Team on Optimization of Logistic systems. He is an IEEE Senior Member and specializes in distributed optimization, multi-agent systems, supply chain management and metaheuristics. Mekki Ksouri is Professor and Head of the Systems Analysis, Conception and Control Laboratory at Tunis El Manar University, National Engineering School of Tunis (ENIT) in Tunisia. He is an IEEE Senior Member and specializes in control systems, nonlinear systems, adaptive control and optimization.