A hybrid sine cosine algorithm with pattern search (SCAPS) for enhanced digital IIR filter system identification

This research introduces a sine cosine algorithm with pattern search (SCAPS), a hybrid optimization algorithm for digital infinite impulse response (IIR) filter system identification. SCAPS combines the global exploration of SCA with the local exploitation of pattern search, addressing SCA’s limitations in complex optimization landscapes. It was benchmarked against SCA, genetic algorithm, particle swarm optimization, and cooperation search algorithm. Performance was evaluated using fifth-order IIR plant models and reduced-order fifth- and sixth-order systems representing diverse dynamics. A mean square error (MSE)-based fitness function was used, with statistical metrics (best, worst, mean, standard deviation) and Wilcoxon’s rank-sum test for evaluation. Sensitivity and computational cost analyses were also conducted. Results show that SCAPS achieves faster convergence, lower MSE, and improved robustness, demonstrating its potential for accurate system identification and engineering applications requiring precise parameter estimation.