Doctorado en Ingeniería Matemática · 2022
Vehicle routing optimization in bicycle sharing systems
Full text 205 pages with text out of 212
Read the full thesis Record in the repository
Contents
- Destroy operatorsp. 118
- A MILP as repair operatorp. 119
- An enumeration algorithm for solution repairingp. 123
- General structurep. 99
- Dominance rules for partial solutionsp. 126
- Computational experimentsp. 69
- Instances and experiments configurationp. 127
- Results on mixed integer lineal modelp. 128
- Comments on multi-start iterative LNS matheuristic performancep. 136
- Comments on enumeration algorithm for ALNS-based strategiesp. 139
- Concluding remarksp. 49
- Introduction and motivationp. 146
- A brief review on two-echelon routing problemsp. 148
- Mathematical model for the 2E–BRPSDp. 150
- Hybrid constructive algorithm for the 2E–BRPSDp. 154
- Greedy randomized constructionp. 62
- Variable neighborhood descentp. 17
- Splitp. 158
- Local searchp. 158
- Global feasibility verificationp. 162
- A set partitioning problem based matheuristicp. 163
- Sets of central and secondary routes for the 2E–BRPSDp. 163
- Set partitioning problem for secondary routesp. 167
- An enhanced GTSP mathematical formulationp. 168
- Computational experimentsp. 69
- Data setsp. 69
- Results on 2E–BRPSD MILPp. 173
- Results on matheuristic algorithmsp. 173
- Results on EnCicla instancesp. 186
- Concluding remarksp. 49
- Removing adjacent nodes to a random arcp. 119
- Destroy methods weight for different sizes of 1–PDVRPLCp. 139
- Goals weight for different sizes of 1–PDVRPLCp. 140
- An example of a 2E–BRPSD solution for Encicla in Medell´ın (Colombia)p. 147
- Optimal solution for a 2E–BRPSD instance with 20 stations and four vehiclesp. 153
- Example for route index computationp. 170
- Results on 1–PDVRPLC instancesp. 129
- Average cost increase based on α = |K|−1p. 130
- Average CPU time (s) for the 1-PDVRPRL MILPp. 130
- Comparative results on LNS-based algorithms for instances with |N| = 20p. 132
- Comparative results on LNS-based algorithms for instances with |N| = 30p. 133
- Comparative results on LNS-based algorithms for instances with |N| = 40p. 134
- Comparative results on LNS-based algorithms for instances with |N| ≥50p. 135
- Comparative results on CPU times (s) for LNS-based algorithmsp. 138
- 4.10 Results on LNS procedure and SPP IP for the MS-ILNSp. 138
- 4.11 Repair MILP in ALNS-Math vs. EA on K′p. 141
- 4.12 Repair MILP in ALNS-Math vs. EA on Kp. 141
- Values for vehicles load when visiting central route stationsp. 154
- Feasible path for a secondary routep. 160
- Number of vehicles for 2E–BRPSD instancesp. 173
- MILP results for total traveling cost on 2E–BRPSD instancesp. 174
- MILP gaps and CPU times (s) for 2E–BRPSD instancesp. 175
- 5.13 CPU times for 2E–BRPSD matheuristic strategies (s)p. 184
- 5.14 CPU time distribution for hybrid constructive algorithm componentsp. 184
- 5.15 CPU time distribution for SPP-based algorithm componentsp. 184
- 5.17 MILP results on EnCicla instancesp. 187
- 4.11 Repair solution algorithmp. 125
- Hybrid constructive algorithm for the 2E–BRPSDp. 154
- SPP based algorithm for the 2E–BRPSDp. 163
- GTSP-based algorithm for the 2E–BRPSDp. 166
Keywords
Bike sharing systemsMathematical programmingMetaheuristicsPickup and deliveryMetaheurísticasOptimizacionRuteo de vehículosSistemas de bicicletas compartidas