A MODIFIED PARTICLE SWARM OPTIMIZATION WITH MUTATION AND REPOSITION

The common problems of particle swarm optimization (PSO) are those of trapping in local optimum and premature convergence. This research paper aims to develop a solution to both problems by introducing mutation around particles and employing the reposition technique. The concurrent use of the introduced mutation and reposition has proved to solve both problems and enhanced the PSO performance; and thus is employed in this research. The proposed technique is termed MRPSO. MRPSO is tested on sixteen benchmark functions and the multidimensional knapsack problems (MKP). MRPSO yields the more satisfactory search results than the genetic algorithm (GA) and PSOs for the benchmark functions and the MKPs.