A vehicle routing problem with backhauls and time windows: a guided local search solution

This paper presents a guided local search heuristic to solve a vehicle routing problem with backhauls and time windows. The VRPBTW with and without customer precedence are both considered. Customer precedence requires that all linehaul customers be visited before any backhaul customer. The basic approach is to construct an initial infeasible solution and then use a guided local search to improve the solution feasibility and quality. A new technique--section planning--is used to enhance the feasibility. Computational results show that the new heuristic can solve problems in which customers are distributed normally or in clusters. Some of the results are better than the best solutions that have appeared in the literature.