In this paper we present a Scatter Search algorithm designed for the solution of the Classical Vehicle Routing Problem, giving details of each of its basic phases when applied to the routing problem. Scatter Search is an evolutionary meta-heuristic that combines solutions with the objective of obtaining new solutions of higher quality; even though there are similarities between Scatter Search and Genetic Algorithms, the two methods differ in basic principles. An important aspect of Scatter Search is to construct high quality solutions in order to direct the search to promising regions. Computational experiments were conducted using four data sets available in the literature. The results show that Scatter Search is robust and competitive in terms of both the quality of solutions and computational times, for the VRP data sets we used.
scatter search; vehicle routing problem; meta-heuristics
