Improved Northern Goshawk Optimization Algorithm for Global Optimization

Global optimization has been used in many real-world problems. Nature-inspired meta-heuristic algorithms, such as the Northern Goshawk Optimization NGO algorithm that has just been proposed, are often used to solve these kinds of optimization problems. An NGO provides satisfactory results. In this algorithm, the proposed exploration model may not provide sufficient coverage of the problem space, trapping the system in a local optimal solution. To improve the performance of NGO, a novel and efficient improved northern goshawk optimization technique named INGO is proposed in this paper. In INGO, a new concept of switching between exploration and exploitation has been developed to improve overall algorithm performance to avoid being stuck in local optima. Also, to increase search capabilities, Levy Flight is used. Twenty-three known benchmark functions were used to test the performance of the proposed INGO. The results were compared to those of an NGO and some well-known robust algorithms. Experimental data indicates that the INGO suggested in this study consistently outperforms the traditional NGO and alternative methods in a significant number of test functions.