Novel Technique for PID Tuning by Particle Swarm Optimization

An attempt has been made by incorporating some special features in the conventional particle swarm optimization (PSO) technique and its usefulness was tested in a common control application involving PID controller tuning. This algorithm differs from the existing PSO methods in the inclusion of unbiased search approach of the swarm particles without establishing inter-particle communication in the initial phase itself, built-in acceleration mechanism of the particles trapped in the local minima of the objective function, preserving the initial population without reproducing those with favourable costs by cluster unification and finer search phase to identify the global minimum. The results of the aforementioned application in comparison with practical methods are quite encouraging. Index terms – population, swarm particle, local minimum, cost, Euclidean distance.