Solving the stochastic location-routing problem with genetic algorithm

Location-routing Problem is a kind of hard combinatorial optimization problem arose in supply chain and logistics system. The deterministic location-routing problems in which all data are known in advance have been researched sufficiently, and the conventional method of solving the deterministic problems is to divide the, problems into location-allocation problems and vehicle routing problems. The stochastic location-routing problem which is much closer to the real case is researched in this paper. In the stochastic problem the demands of the customers follow a certain random distribution. A genetic algorithm is designed to solve the stochastic location-routing problem. Novel genetic represent and corresponding genetic operations are designed in the genetic algorithm so that the location-allocation and vehicle-routing can be tackled simultaneously. Local search is also applied in the algorithm in order to improve the search effectiveness and the solution quality. Simulations based on numerical examples show that the proposed algorithm is effective.