A Novel Real Number Encoding Method of Particle Swarm Optimization for Vehicle Routing Problem

Abstract

Vehicle routing problem is a well-known NP problem, many heuristic algorithms, such as genetic algorithm, simulated annealing algorithm is applied in the problem. Particle swarm optimization (PSO) is a new evolutionary computation technique. Although PSO algorithm possesses many attractive properties, the method of encoding in NP problem need further to investigated. In the paper, a novel real number encoding method of particle swarm optimization (PSO) for vehicle routing problem is proposed. The vehicle is mapped into the integer part of the real number; the sequence of the customers in the vehicle is mapped into the decimal fraction of the real number. After decoding, saving algorithm, nearest neighbor algorithm and or-opt optimizes the inner or outer routes and modify the illegal solution. Series of numerical examples were tested and verified, which shows the better performance of the proposed algorithm compare with other particle swarm optimization algorithm and genetic algorithm