This book reports on an operational management approach to improving bike-sharing systems by compensating for fluctuating demand patterns. The aim is to redistribute bikes within the system, allowing it to be “actively” balanced. The book describes a mathematical model, as well as data-driven and simulation-based approaches. Further, it shows how these elements can be combined in a decision-making support system for service providers. In closing, the book uses real-world data to evaluate the method developed and demonstrates that it can successfully anticipate changes in demand, thus supporting efficient scheduling of transport vehicles to manually relocate bikes between stations.
Author(s): Jan Brinkmann
Series: Lecture Notes In Mobility
Edition: 1st Edition
Publisher: Springer
Year: 2020
Language: English
Pages: 195
Tags: Transportation Technology And Traffic Engineering
Foreword......Page 6
Preface......Page 8
Contents......Page 9
Acronyms......Page 12
List of Figures......Page 13
List of Tables......Page 15
List of Algorithms......Page 19
1 Introduction......Page 20
Part I Preliminaries......Page 24
2.1 Urban Mobility......Page 25
2.2.1 Reduction of Traffic......Page 26
2.2.3 Increase in Tourists Attractiveness......Page 27
2.3.2 Station-Based......Page 28
2.4 Request Patterns......Page 29
2.4.3 Leisure and Tourists......Page 30
2.5.1 Strategical Management......Page 31
2.5.2 Tactical Management......Page 33
2.5.3 Operational Management......Page 35
3.1.1 Traveling Salesman Problem......Page 37
3.1.5 Inventory Routing Problem......Page 38
3.2 Inventory Routing for Bike Sharing Systems......Page 39
3.2.1 No Request......Page 40
3.2.2 Request......Page 41
4.1 Markov Decision Processes......Page 48
4.2 Approximate Dynamic Programming......Page 52
4.2.2 Lookahead......Page 53
4.2.3 Value Function Approximation......Page 55
Part II Application......Page 58
5.1 Narrative......Page 59
5.3 Markov Decision Process......Page 60
5.4 Example......Page 63
5.5 Challenges......Page 64
6.1 Outline......Page 67
6.2.1 Simulation......Page 69
6.2.2 Optimization......Page 73
6.3.1 Lookahead Policy......Page 80
6.3.2 Online Simulations......Page 81
6.3.3 Offline Simulations......Page 82
6.3.4 Matrix Maximum Algorithm......Page 83
7.1 Outline......Page 84
7.2 Definition......Page 86
7.2.1 Sequences of Lookahead Horizons......Page 87
7.2.2 Value Function Approximation......Page 88
7.2.3 Boltzmann Exploration......Page 89
7.3.1 Value Function Approximation......Page 92
7.3.2 Boltzmann Exploration......Page 94
8.1.1 Data Preprocessing......Page 95
8.1.2 Resulting Data Set......Page 96
8.2 Instances......Page 97
8.3 Transition......Page 98
8.4 Benchmarks......Page 99
8.4.1 Safety Buffer-Tending Relocation Policy......Page 101
8.5 Parametrization......Page 102
8.5.2 Online Simulations......Page 103
8.5.4 Dynamic Lookahead Policies......Page 105
8.5.5 Rollout Algorithms......Page 106
8.6 Results......Page 108
8.6.2 The Value of Anticipation......Page 109
8.6.3 Individual Results......Page 110
8.7.1 Optimal Assignment......Page 112
8.7.2 Learning Curves......Page 114
8.7.3 Dynamic Lookahead Horizons......Page 115
Part III Conclusion......Page 120
9 Managerial Implications......Page 121
10.1 Model......Page 123
10.2 Method......Page 126
Appendix A Parameters......Page 129
Appendix B Results......Page 136
BookmarkTitle:......Page 188
