Tuning a Parametric Clarke-Wright Heuristic for Vehicle Routing Through a Genetic Algorithm

Almost all heuristic optimization procedures require the presence of a well tuned set of parameters. The tuning of these parameters is usually a critical issue and may entail intensive computational requirements. In our work we propose a fast and effective approach composed of two distinct stages. In the first stage a genetic algorithm is applied to a small subset of representative problems in order to determine few robust parameter sets. In the second stage these sets of parameters are the starting points for a fast local search procedure, able to investigate deeper the space of parameter sets for each problem to be solved. This method is tested on a parametric version of the Clarke and Wright algorithm and the results are compared with the enumerative approach previously considered by the algorithm’s authors. Our method is able to produce results of similar quality respect the enumerative approach, but requires much shorter computational time.