A Novel Hybrid Multi-Population GWO-BWO Method for IIR Filter Parameter Estimation

Abstract

This study presents a novel evolutionary optimization technique known as the hybrid multi-population evolutionary algorithm, which blends the Grey Wolf Optimizer (GWO) and Black Widow Optimization (BWO) methodologies. The method is driven by the drawbacks of single-population algorithms, namely their inability to handle local optima and premature convergence. The suggested hybrid algorithm improves convergence performance by fusing BWO's fierce local search and cannibalism technique with GWO's global exploration capacity. The approach's Mean Square Error (MSE), convergence speed, and robustness are assessed while being tested against benchmark IIR filters. Experimental findings validate the superiority of the hybrid GWO-BWO technique when compared to Flower Pollination Algorithm (FPA), Particle Swarm Optimization (PSO), and Genetic Algorithm (GA).