Particle Swarm Optimization for the Orienteering Problem

In this paper, we propose a Particle Swarm Optimization (PSO) based solution to the Orienteering Problem. A major problem in many PSO applications is the premature convergence of some particles to local optimums. To overcome this problem in our application, such particles are re-initiated; consequently, the exploration capability of PSO is improved. Furthermore, by employing Reduced Variable Neighborhood Search as a local search method for these particles, we enhanced exploitation mechanism. By these modifications, PSO achieved best known solutions by evaluating smaller part of the search space than similar results in literature.