IZOA: Multi-strategy improved zebra optimization algorithm and its engineering applications

Abstract

The Zebra Optimization Algorithm (ZOA) mimics the social behavior of zebras and is susceptible to the interference of local optimal solutions, leading to poor optimization and premature convergence. In this paper, we propose an improved zebra optimization algorithm (IZOA) that integrates several advanced strategies to overcome these problems. First, IZOA introduces a Lévy flight strategy in the foraging phase of the zebra population to expand the search range and improve the quality of individuals. At the same time, the “PZ” mechanism updates the other individuals based on the value of the leading zebra in each generation, which accelerates the optimization process and improves the searching ability. In addition, IZOA integrates a nonlinear convergence factor based on the COS function, which improves the convergence speed and balances the exploration and development phases. A Cauchy variation strategy is used to enhance the global search capability and help the population escape from local extremes. In CEC2017 and CEC2022 benchmarking and rolling bearing design applications, IZOA is compared with 12 mainstream and improved ZOA algorithms (CZOA and IIZOA), and shows better performance. Finally, IZOA is combined with LSTM network for wind power prediction to show its application advantages in real engineering design problems.