A Mixed PSO algorithm for the VRPSPD

This paper investigates the reverse logistics vehicle routing problem with a single depot, simultaneous distribution and collection of the goods by a homogeneous fleet of vehicles under the restrictions of maximum capacities and maximum distance. A mixed integer programming model is constructed To solve this problem, a mixed PSO (particle swarm optimization and simulated annealing) algorithm is designed This paper designs the coding rules of initial solution, 2-opt information exchange strategy, neighborhood solution generation strategy and cooling schedule of SA algorithm strategy. The penalty function is used in the fitness function to avoid the generation of infeasible solutions. Taking one typical instances as example, the experimental study indicates that this approach could improve the vehicle load rate and get rid of the additional total distance caused by the fluctuating vehicle load and the limited capacity. The satisfying solutions can be obtained in acceptable computational time.