An Improved Discrete Particle Swarm Optimization Algorithm and its Application

摘要

Evolutionary algorithms solve complex identification problems because they do not take into account any constraint on the cost function and so more flexibility is offered when the model structure is chosen and the cost function is minimized.Step input based identification has been developed,which could be used to determine the transient behavior of a system rapidly.In this paper,a new DPSO algorithm presented and the performance of this algorithm for solving identification problems compared with the GA.The obtained results demonstrate that the proposed DPSO algorithm performs rather well in terms of speed (FNGC) and reaching to the minimum cost.