Solve Capacitated Vehicle Routing Problem Using Hybrid Chaotic Particle Swarm Optimization

Vehicle Routing Problem (VRP) is a classical NP-hard problem in combinational optimization with great practicality. A branch of VRP is the Capacitated Vehicle Routing Problem (CVRP), where vehicles have capacity constraints. In this paper, a Hybrid Chaotic Particle Swarm Optimization (HCPSO) is proposed to solve CVRP. To make this algorithm successful, three main components play an important role. Firstly, we introduce a new mutual mapping method to encode and decode between decimal numbers and integer solutions. Secondly, the ergodicity and sensitivity on initial conditions of chaos theory are utilized to achieve chaotic initialization and renewing. Thirdly, various local search strategies like neighbor change strategy, move strategy are employed to improve the local search ability. In the end, benchmarks have been tested. It is clearly shown that for small and medium sized CVRP, the algorithm is efficient and effective and the results are better than those of some recent papers.