Solving Capacitated Vehicle Routing Problems via Genetic Particle Swarm Optimization

A modified genetic particle swarm optimization method (MGPSO) is employed to solve the capacitated vehicle routing problems (CVRP). MGPSO was derived from the standard particle swarm optimization (PSO) and incorporated with the genetic reproduction mechanisms, namely crossover and mutation. MGPSO employs an integer encoding and decoding representation, which is suitable for combinatorial optimization problems and it is easy to be implemented. Moreover, with the encoding and decoding, it can adjust the number of vehicles needed, dynamically and adaptively. A modified ordering crossover is employed to perform the crossover based on the PSO mechanisms to exchange building blocks with personal and global experiences. The proposed method has been implemented to five well-known CVRP benchmarks, and by comparison with the other evolutionary algorithms, the simulation results have shown the feasibility and effectiveness.