PSO Algorithm Based on Accumulation Effect and Mutation

Abstract

Particle SwarmOptimization (PSO) algorithm is a new swarm intelligence optimization technique, because of its simplicity, fewerparameters and good effects, PSO has been widely used to solve various complexoptimization problems. particle swarm optimization(PSO) exist the problems ofpremature and local convergence, we proposed an improved particle swarm optimization based on aggregation effect and with mutation operator, whichdetermines whether the aggregation occurs in searching, if there is then theGaussian  mutation is detected to theglobal extremum, to overcome particle swarm optimization falling into localoptimal solution defects.  Testing thenew algorithm by a typical test function, the results show that, compared withthe conventional genetic algorithm (SGA), it improves the ability of globaloptimization, but also effectively avoid the premature convergence.

DOI: http://dx.doi.org/10.11591/telkomnika.v11i12.3703