Adaptive Particle Swarm Optimization for low pass finite impulse response filter design

This paper presents an optimal design of linear phase digital finite impulse response (FIR) low pass (LP) filter using Adaptive Particle Swarm Optimization (APSO). APSO is an improved version of conventional PSO in which estimation of evolution state determines one of the four phases such as exploration, exploitation, convergence and jump-out for the entire optimization process. These modifications bring better balance in exploration and exploitation stages in multidimensional search space that leads to near global optimal solution. A comparison of simulation results reveals the optimization efficacy of APSO over the prevailing optimization techniques like real coded genetic algorithm (RGA) and PSO for the solution of the multimodal, non-differentiable, highly non-linear., and constrained FIR filter design problems.