A hybrid Particle Swarm Optimization algorithm for combinatorial optimization problems

Particle Swarm Optimization (PSO) belongs to a class of algorithms inspired by natural social intelligent behaviors, called Swarm Intelligence (SI). PSO has been successfully applied to solve continuous optimization problems, however, its potential in discrete problems has not been sufficiently explored. Recent works have proposed hybridization of PSO using local search and Path relinking algorithms with promising results. This paper aims to present a hybrid PSO algorithm that uses local search and Path relinking too, but differently to the previous approaches, this works maintains the main PSO concept for the update of the velocity of the particle. The paper describes the proposed algorithm and a set of experiments with the Traveling Salesman Problem (TSP). The results are compared to other Particle Swarm Optimization algorithms presented previously for the same problem. The results are encouraging and reinforce the idea that PSO algorithms can also provide good results when dealing with discrete problems.