Abstract: In order to solve the multimodal optimization problem, a Deterministic Crowding based Coyote Optimization Algorithm (DCCOA) is proposed. Deterministic crowding, which is a niching technology, is integrated into the Coyote Optimization Algorithm (COA) to propose a new coyote evolution mechanism. Meanwhile, the calculation method of the cultural trend of the coyote pack is improved. In addition, the weight method is used to update the social conditions of coyote for simulating the population life of coyote more realistically. Experiments on typical benchmark functions with different dimensions of decision variables are carried out by the DCCOA and other intelligent optimization algorithms. The experimental results show that the niching technology employed in the DCCOA further promotes the COA balance between exploration and exploitation and improves the global optimization ability of the COA in multimodal situation. The proposed algorithm improves the convergence accuracy, convergence speed and stability.

Key words: coyote optimization algorithm, multimodal optimization, deterministic crowding, niching