关键词: 多车场多车型车辆调度问题, 粒子群算法, 柯西变异

Abstract: Multi-Depots and Multi-Vehicles Vehicle Scheduling Problem（MDMVVSP） is a kind of NP combination problem in logistics distribution. In order to overcome the problems such as long computing time and easy to fall into local best for traditional Particle Swarm Optimization（PSO）, an Improved PSO（IPSO） algorithm is put forward. In the algorithm, Cauchy mutations are used to ensure lager mutation steps and escape from local optimal solution. The algorithm is applied to MDMVVSP, the simulation results show that the algorithm is feasible to solve the MDMVVSP, and it is better than PSO.