Elephant herding optimization based on Orthogonal opposition-based learning and Sugeno inertia weights

摘要

In the engineering field, many global optimization problems are nonlinear, multimodal, and complex. Elephant herding optimization performs well in solving these problems. The original elephant herding optimization suffers from several shortcomings, such as low convergence accuracy and eases to fall into the local optimum. In this paper, an elephant herding optimization based on orthogonal opposition-based learning and Sugeno inertia weights, called EHOOBS, is proposed. First, orthogonal-opposition learning is embedded into the proposed algorithm to improve the optimization seeking efficiency. By designing orthogonal tests to construct candidate solutions in some dimensions, the information of forward and reverse agents in different dimensions is fully utilized. Furthermore, the optimal balance between exploration and exploitation is tried to provide by introducing Sugeno inertia weight into the proposed method. Based on the modifications, better solutions are presented for the algorithm. To verify the effectiveness of the algorithm, experiments are conducted on the benchmark function. The results show that the proposed algorithm has stronger convergence performance and superior capabilities of exploration and exploitation.