Particle swarm optimization for multi-depots single vehicle routing problem

Abstract

A multi-depots single vehicle routing problem (VRP) model based on goods were established to get global solution for VRP. Particle position matrix based on goods was established for particle swarm optimization (PSO) to sovle VRP model, and every matrix column corresponded to one goods. Matrix had two rows, first row corresponded to goods start point, and second row corresponded to goods end point. Matrix elements were random number between 0 and 1, and matrix elements sort rules were established to get single vehicle route. Then evaluated and filtered particles by objective function until met terminate condition. In this VRP model strategy, vehicle route to all goods was searched to avoid local minima. Illustration results show that VRP model and algorithm are effective and offer a way to multi-depots single vehicle routing problem.