A self-tuning heuristic for a multi-objective vehicle routing problem

In this study, a heuristic free from parameter tuning is introduced to solve the vehicle routing problem (VRP) with two conflicting objectives. The problem which has been presented is the designing of optimal routes: minimizing both the number of vehicles and the maximum route length. This problem, even in the case of its single objective form, is NP–hard. The proposed self–tuning heuristic (STH) is based on local search and has two parameters which are updated dynamically throughout the search process. The most important advantage of the algorithm is the application convenience for the end–users. STH is tested on the instances of a multi–objective problem in school bus routing and classical vehicle routing. Computational experiments, when compared with the prior approaches proposed for the multi–objective routing of school buses problem, confirm the effectiveness of STH. STH also finds high–quality solutions for multi–objective VRPs.