An oppositional wolf pack algorithm for Parameter identification of the chaotic systems

Abstract Parameter identification is an important issue for chaotic systems, which have been investigated intensely in several fields such as secure communication, power converters and biological systems. To solve the above problem, most existing evolutionary algorithms were proposed. However, the existing evolutionary algorithms have their own limitations when solving the parameter identification problem. Therefore, we propose a novel and efficient oppositional wolf pack algorithm (OWPA), which has good balance of exploitation and exploration, to estimate the parameters of Lorenz chaotic system. First, an oppositional initial population producing method is proposed to enhance the global convergence. Second, the opposition-based learning method is presented for new wolf producing in wolf pack algorithm, which cannot only enhance the local searching ability of wolf pack algorithm in normal situation but also increase the diversity of distributed population in catastrophic situation. Furthermore, the proposed oppositional wolf pack algorithm is applied to estimate the parameters for Lorenz chaotic system under the offline and online condition. Finally, numerical example is provided to illustrate the effectiveness of the proposed algorithm. It is shown that the above proposed algorithm has more effectiveness and robustness than the existing popular algorithms (i.e., particle swarm optimization algorithm, firefly algorithm and cuckoo search algorithm).