Research on Optimization Algorithm and Application of Improved Northern Goshawk Based on Fusion Strategy

ABSTRACT Northern Goshawk Optimization Algorithm (NGO), as a new swarm intelligence optimization algorithm, shows certain potential in solving complex optimization problems because of its unique search mechanism. However, the algorithm still faces some challenges in practical application, such as slow convergence speed, insufficient optimization precision, and easy to fall into local optimality. These problems limit its application range and efficiency in complex multimodal optimization problems. To overcome the above shortcomings, this paper proposes an improved Northern Goshawk optimization algorithm (WNGO) based on a fusion strategy. The fusion strategy is a novel approach that combines the strengths of different optimization algorithms to address the problems of slow convergence speed, accuracy of optimization, and easy falling into local optimal. First, the Piecewise chaotic mapping is used to initialize the Northern Goshawk population, which enhances the global search capability of the algorithm by providing a wider search space in the initial stage. Second, in order to achieve the adequacy of the solution space search and the performance of the optimization problem in the prey recognition stage of the Northern Goshawk, the location update formula of the prey recognition stage of the Northern Goshawk is replaced by the location update formula of the Walrus optimization algorithm in the exploration stage. Then, through the mirror reverse learning strategy, the reverse solution generated by the lens imaging principle can provide a new search direction when the Northern Goshawk optimization algorithm falls into the local optimal, increase the probability of finding the global optimal solution, and improve the global optimization ability, so that it can jump out of the local optimal in the later iteration. Finally, the adaptive T‐distribution variation strategy is used to enhance the local exploration ability in the late iteration, thus improving the convergence speed of the Northern Goshawk optimization algorithm. This paper evaluates the performance of the improved WNGO algorithm. By comparing the CEC2021 test function and other advanced improved swarm intelligence methods, it is proved that the improved algorithm has better accuracy, robustness, and convergence speed. It is tested in two engineering design problems. The results show that the WNGO algorithm can break through the local optimal solution, obtain higher precision, and have a stronger global searching ability than other algorithms.