A path-exchange-type local search algorithm for vehicle routing and its efficient search strategy

This paper proposes a fast path-exchange-type local search algorithm for vehicle routing problems (VRPs) that uses a new search strategy to restrict the search to promising neighbors. An efficient search strategy for path exchange that uses edge lengths to give search radii is proposed. Intensive numerical experiments with VRP instances of up to 10,000 vertices were conducted, and the results are compared with those given by the k-Opt algorithms. It is shown that our algorithm is faster than the 3-Opt and yields the quality of the solution comparable with the 4-Opt.