Double-Swarm Anti-Moderation Particle Swarm OptimizationAlgorithm

Particle swarm optimization (PSO) has been adopted to solve myriad ofoptimization problems in different fields of science and engineering. In thispaper a new Double-Swarm PSO algorithm has been introduced which notonly has a fast convergence speed but also exhibits excellent performance infinding optimum solutions. This method differs from basic PSO algorithm intwo aspects. Firstly, a negative component named anti-moderation term hasbeen added to velocity vector of each particle. This term is calculated byaveraging the positions of current particle and global best. In fact, this policyhas been proposed in order to stimulate the particles toward unexplored areasinstead of just around global best position. The other feature of method isclustering the swarm into two separate randomly generated sub-swarms. Thisfeature significantly speeds up the convergence of algorithm whilemaintaining the performance of results. The algorithm has been tested in caseof various benchmark problems. Its performance also has been compared withsome meta-heuristic algorithms and variants of PSO. The results show therelative superiority of method among analyzed methods both in terms of speedand performance

کلیدواژه ها:

Anti-Moderation Velocity Component ، Particle Swarm Optimization (PSO) ، Double-Swarm PSO