AI SearchAdaptive variable neighborhood search algorithms for solving capacitated vehicle routing problems
Sikarin Vinyoopradit
Adaptive variable neighborhood search algorithms for solving capacitated vehicle routing problems - Pathum Thani : Asian Institute of Technology, 2012 - 104 p. : ill. (some col.), charts - Thesis ; no. ISE-12-51 . - Asian Institute of Technology. Thesis ; no. ISE-12-51 .
Submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Industrial and Manufacturing Engineering, School of Engineering and Technology
Thesis (M. Eng.) - Asian Institute of Technology, 2012
This research presents adaptive variable neighborhood search algorithms for solving Capacitated Vehicle Routing Problems (CVRP). Four Variable Neighborhood Search (VNS) based algorithms to optimize Capacitated Vehicle Routing Problems (CVRP) are studied. The first algorithm is a Variable Neighborhood Search (VNS) integrated with Tabu Search (TS) implemented with new proposed Tabu List. The second algorithm is an Adaptive Variable Neighborhood Search (AVNS) algorithm to improve the perturbation process of VNS algorithm by allowing the frequently used sequences in local optimum to be selected via integrating Roulette Wheel Selection method. The third algorithm is an alternative AVNS algorithm that adopts approach similar to Ant Colony Optimization (ACO) to the AVNS algorithm called A-AVNS1 to reduce the influence of the infrequent used sequences. An Adaptive Roulette Wheel Selection method is proposed in A-AVNS2 algorithm to improve algorithm exploration efficiency. Numerical experiments are carried out using the published benchmark test problems. The results indicated that the proposed adaptive VNS algorithms improved solution quality and solution time for large problem sizes.
Vehicle routing problem
Algorithms
Adaptive variable neighborhood search algorithms for solving capacitated vehicle routing problems - Pathum Thani : Asian Institute of Technology, 2012 - 104 p. : ill. (some col.), charts - Thesis ; no. ISE-12-51 . - Asian Institute of Technology. Thesis ; no. ISE-12-51 .
Submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Industrial and Manufacturing Engineering, School of Engineering and Technology
Thesis (M. Eng.) - Asian Institute of Technology, 2012
This research presents adaptive variable neighborhood search algorithms for solving Capacitated Vehicle Routing Problems (CVRP). Four Variable Neighborhood Search (VNS) based algorithms to optimize Capacitated Vehicle Routing Problems (CVRP) are studied. The first algorithm is a Variable Neighborhood Search (VNS) integrated with Tabu Search (TS) implemented with new proposed Tabu List. The second algorithm is an Adaptive Variable Neighborhood Search (AVNS) algorithm to improve the perturbation process of VNS algorithm by allowing the frequently used sequences in local optimum to be selected via integrating Roulette Wheel Selection method. The third algorithm is an alternative AVNS algorithm that adopts approach similar to Ant Colony Optimization (ACO) to the AVNS algorithm called A-AVNS1 to reduce the influence of the infrequent used sequences. An Adaptive Roulette Wheel Selection method is proposed in A-AVNS2 algorithm to improve algorithm exploration efficiency. Numerical experiments are carried out using the published benchmark test problems. The results indicated that the proposed adaptive VNS algorithms improved solution quality and solution time for large problem sizes.
Vehicle routing problem
Algorithms