A self-adaptive local search algorithm for the classical vehicle routing problem

The purpose of this study is introduction of a local search heuristic free from parameter tuning to solve classical vehicle routing problem (VRP). The VRP can be described as the problem of designing optimal delivery of routes from one depot to a number of customers under the limitations of side constraints to minimize the total traveling cost. The importance of this problem comes from practical as well as theoretical point of view. The proposed heuristic, self-adaptive local search (SALS), has one generic parameter which is learnt throughout the search process. Computational experiments confirm that SALS gives high qualified solutions to the VRP and ensures at least an average performance, in terms of efficiency and effectiveness, on the problem when compared with the recent and sophisticated approaches from the literature. The most important advantage of the proposed heuristic is the application convenience for the end-users. SALS also is flexible that can be easily applied to variations of VRP.