Solving Multi-objective Vehicle Routing Problem Using Hyper-heuristic Method By Considering Balance of Route Distances

Abstract

Vehicle Routing Problem (VRP) is one of the combinatoric problems that is difficult to solve, so it is incorporated into an NP-hard problem. VRP aims to produce a set of shortest routes from several of the same capacity vehicles to visit several customers at a certain time limit. Depot is the starting and ending point of the route. Due to the complexity of industry needs, the VRP problem needs to be improved into a multi-objective. Most of VRP prior researches only minimize total distance as a single objective. Therefore, in this study added an objective related to the balance of distances between routes. In prior researches, multi-objective VRP was solved using metaheuristic. It requires the determination of parameters and specific algorithm design to solve each problem domain. To overcome these shortcomings, this study uses a hyper-heuristic method to complete multi-objective VRP. Given that the use of hyper-heuristics in previous studies is for single objective VRP, so this study also proposes hyper-heuristic for multi-objective VRP. Gehring and Homberger dataset is used for the experiment. Based on the experiments in this study, The Hill Climbing algorithm gives better results than The Great Deluge algorithm for completing multi-objective VRP.