An Analysis of Archive Update for Vector Evaluated Particle Swarm Optimization

Abstract

Multi-objective optimization problem is commonly found in many real world problems. In computational intelligence, Particle Swarm Optimization (PSO) algorithm is a popular method in solving optimization problems. An extended PSO algorithm called Vector Evaluated Particle Swarm Optimization (VEPSO) has been introduced to solve multi-objective optimization problems. VEPSO algorithm requires an archive, which is used to record the solutions found. However, the outcome may be differ depending on how the archive is used. Hence, in this study, the performance of VEPSO algorithm when updates the archive at different instance is investigated by measuring the convergence and diversity by using standard test functions. The results show that the VEPSO algorithm performs better when update the archive during the search process, in the iterations.