A breeder genetic algorithm for vehicle routing problem with stochastic demands

Abstract

This paper considers a version of VRP known as VRP with Stochastic Demands (VRPSD) where the demands are unknown when the route is designed. The problem objective is to find a priori route under preventive restocking that minimize the total expected cost, including travel cost and the expected recourse cost, subject to the routing constraints, under the stochastic demands setting. The Breeder Genetic Algorithm is proposed to solve this problem. BGA is a kind of GAs, which is especially powerful and reliable in global searching. The BGA was compared to the standard Genetic Algorithm on a set of randomly generated problems following some discrete probability distributions. The problem data are inspired by real case of VRPSD in waste collection. From the results, it was found that the BGA was clearly superior to standard GA in terms of solution quality. Compared to Bianchi et al’s GA, the BGA also may lead to a better performance