An Enhanced and Computationally Efficient Cuckoo Search Algorithm for High-Order System Model Reduction

Abstract

Model order reduction (MOR) is a crucial technique for simplifying high-order dynamical systems while preserving their essential characteristics. In this paper, we propose an enhanced version of the Cuckoo Search Algorithm (CSA) designed to achieve more accurate reduced-order models with lower computational cost. The improved algorithm incorporates adaptive Lévy flight adjustments and an optimized nest updating strategy, leading to faster convergence and better stability in complex system approximations. Comparative evaluations against classical CSA and other metaheuristic-based MOR techniques demonstrate the superiority of our approach in terms of accuracy, computational efficiency, and robustness. The proposed method offers a promising solution to reduce computational burden in control and simulation tasks involving high-dimensional systems.