Diversity Preserving Auto Improved – PSO for Solving Optimization Problems
Ashok Kumar, Brajesh Kumar Singh and B.D.K Patro
Particle Swarm Optimization [PSO] is a very popular optimization algorithm and due to its simplicity it has been used in many applications. Although PSO converges very fast yet it has stagnation problem. To improve its convergence rate and to remove stagnation problem, some changes in velocity vector is suggested. These changes motivate each particle of PSO in different directions so that full search space can be covered and better solutions can be captured. Moreover, auto tuning of random parameters are done to remove stagnation problem. This auto improved version is named as AI-PSO algorithm. To check the performance of proposed version, it is compared with state of art algorithms such as PSO-TVAC and basic PSO. Result shows the superiority of the algorithm.
Keywords: PSO, Stagnation, Local Optima, Premature Convergence.