Analysis of swarm intelligence and evolutionary computation techniques in IIR digital filters design

Abstract

Digital filters provide excellent advantages, compared to analog filters, such as better stability and precision. According to the duration/length of the impulse response, digital filters are categorized as Finite-Impulse-Response (FIR) and Infinite-Impulse-Response (IIR) filters. Because the error surface of IIR filters is mostly multimodal, powerful global optimization techniques are preferred for avoid local minima in the filter design process. Artificial Intelligence (AI)-based approaches, Swarm Intelligence (SI) and Evolutionary Computation (EC) techniques are candidate methods to address this problem and to produce desirable solutions. SI is used to model the collective behavior of social swarms in nature, such as ant colonies, honey bees, and bird flocks. The EC is based on the principle of evolution (survival of the fittest). In this paper, a novel index for IIR filter design is introduced (called Indicator of Success) and SI and EC algorithms are tested and evaluated for several numbers of novel and conventional heuristic algorithms. The reduced-order identification of two benchmarked IIR plants are carried out. We analyzed the performance of the proposed algorithms in IIR digital filters design in terms of the reliability, Mean-Square-Error (MSE) and IoS. The results demonstrate the proper and reliable performance of the SI algorithms compared to that achieved by EC algorithms.