Improved particle swarm algorithms for global optimization

Particle swarm optimization algorithm has recently gained much attention in the global optimization research community. As a result, a few variants of the algorithm have been suggested. In this paper, we study the efficiency and robustness of a number of particle swarm optimization algorithms and identify the cause for their slow convergence. We then propose some modifications in the position update rule of particle swarm optimization algorithm in order to make the convergence faster. These modifications result in two new versions of the particle swarm optimization algorithm. A numerical study is carried out using a set of 54 test problems some of which are inspired by practical applications. Results show that the new algorithms are much more robust and efficient than some existing particle swarm optimization algorithms. A comparison of the new algorithms with the differential evolution algorithm is also made.