Adaptive learning particle swarm optimizer-II for global optimization

Abstract

This paper presents an updated version of the adaptive learning particle swarm optimizer (ALPSO), we call it ALPSO-II. In order to improve the performance of ALPSO on multi-modal problems, we introduce several new major features in ALPSO-II: (i) Adding particle's status monitoring mechanism, (ii) controlling the number of particles that learn from the global best position, and (iii) updating two of the four learning operators used in ALPSO. To test the performance of ALPSO-II, we choose a set of 27 test problems, including un-rotated, shifted, rotated, rotated shifted, and composition functions in comparison of the ALPSO algorithm as well as several state-of-the-art variant PSO algorithms. The experimental results show that ALPSO-II has a great improvement of the ALPSO algorithm, it also outperforms the other peer algorithms on most test problems in terms of both the convergence speed and solution accuracy.