Decomposition-based multi-objective evolutionary algorithm for vehicle routing problem with stochastic demands

作     者：Gee, Sen Bong Arokiasami, Willson Amalraj Jiang, Jing Tan, Kay Chen 

作者机构：Natl Univ Singapore 4 Engn Dr 3 Singapore 117576 Singapore 

出 版 物：《SOFT COMPUTING》 (Soft Comput.)

年 卷 期：2016年第20卷第9期

页      面：3443-3453页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Ministry of Education - Singapore  MOE  (R-263-000-A12-112) 

主　　题：Vehicle routing problem Multi-objective optimization Evolutionary algorithm 

摘      要：Vehicle routing problem with stochastic demands (VRPSD) is a famous and challenging optimization problem which is similar to many real world problems. To resemble the real world scenario, total traveling distance, total driver remuneration, the number of vehicles used and the difference between driver remuneration are considered and formulated in the multi-objective optimization perspective. This paper aims to solve multi-objective VRPSD under the constraints of available time window and vehicle capacity using decomposition-based multi-objective evolutionary algorithm (MOEA/D) with diversity-loss-based selection method incorporates with local search and multi-mode mutation heuristics. We have also compared the optimization performance of the decomposition-based approach with the domination-based approach to study the difference between these two well-known evolutionary multi-objective algorithm frameworks. The simulation results have showed that the decomposition-based approach with diversity-loss-based selection method is able to maintain diverse output solutions.