Modification Comparisons of the Particle Swarm and Levy Flight Firefly Adaptive DSP Algorithms

Abstract

Previous research results have demonstrated that the bio-inspired Lévy Flight Firefly Algorithm (LFFA) can be effectively used in IIR adaptive filters, non-linear adaptive filters, IIR coupled form adaptive filters, and IIR lattice-ladder adaptive filters. It has recently been shown that the LFFA can be applied to adaptive 2-D McClellan "unconstrained" Transform filters so the adaptivity can approximate the frequency domain contours. This paper demonstrates how a special block length modification to the LFFA algorithm produces a 2-D Modified LFFA (2-D MLFFA) that enhances the adaptive convergence rate and lowers the MSE as the 2D-MLFFA converges toward the global minimum MSE.