Design of IIR filters by determining particle reallocation space using multi-swarm PSO

In this paper, we study a novel method to avoid a premature convergence in the design of IIR (Infinite Impulse Response) filters using PSO (Particle Swarm Optimization). Although PSO is applicable to solve nonlinear optimization problems, it is reported that a local minimum stagnation occurs due to a strong directivity in the searching. Therefore, a method for avoiding the local minimum stagnation plays an important role. In this paper, a method by reallocating particles is proposed, in which multiple swarms are used to determine a reallocation space based on those global bests. Then, the reallocation space is chosen using the stagnation swarm and the another two swarms selected randomly among multiple swarms. Several design examples are shown to verify the effectiveness of the proposed method.