An empirical study of a pure genetic algorithm to solve the capacitated vehicle routing problem

Despite some successful applications of genetic algorithms (GAs) for solving the capacitated vehicle routing problems, they usually rely on problem-specific components or hybridization with other heuristics for obtaining competitive solutions. These problemspecific and hybridized characteristics complicate the GA implementation for solving reallife problems. In this paper, a pure GA for the capacitated vehicle routing problem is proposed. The performance of the algorithm is improved through using a new replacement strategy, which is simple and not problem specific. Computational results show that the new GA solves the problem efficiently and achieves much better results than another pure GA in literature.