Multi-particles Learning Intelligent Swarm Optimizer

Abstract

Based on the conventional Intelligent Single Particle Optimizer(ISPO), learning the multi-swarm in the swarm intelligence algorithm, the paper proposes Multi-particles Learning Intelligent Swarm Optimizer(MLISO). In the optimization process, MLISO randomly generates multi-particles, according to the evolutionary rule of ISPO optimization, after sequence and cycle updating each particle to a certain number of iterations, multi-particles are sorted according to fitness value from good to bad, the individual with worst fitness value would learn from the best individual, if the individual with worst fitness value is still not improved after learning from the best individual with a certain number of times, the new individual is generated randomly again. The MLISO repeats the above steps until the termination condition is reached. Experimental results demonstrates that MLISO, whose optimization capacity exceeded the international improved algorithm based on PSO and ISPO , has some advantages in optimizing complex high-dimensional multimodal function with a large number of local optimal point.