Efficient Improved Henry Gas Solubility Optimization and Its Application in Feature Selection Problems

-The henry gas solubility optimization algorithm is a meta-heuristic algorithm inspired by Henry's law. While it has demonstrated effectiveness in solving various optimization problems, it does face certain limitations such as insufficient population diversity, and slow convergence speed when dealing with complex problems. In this paper, we propose an enhanced version of henry gas solubility optimization algorithm, known as E_HGSO. First, we introduce a new group search formula to improve the ability of avoiding easy to fall into local optimum and searching in a single range, while we introduce the concept of a search factor to strike a balance between exploration and exploitation. Second, we introduce a position update formula to enhance the diversity and randomness of the search process. Finally, we propose a new worst gas position update formula with a L & eacute;vy flight mechanism. This mechanism enhances the gas search's ability to adapt to different distance requirements within the search space, leading to improved search efficiency and accuracy. To evaluate the effectiveness of the E_HGSO algorithm, we conducted a comparison with eight algorithms on the CEC2017 benchmark functions. The results of the Friedman test and Wilcoxon rank sum test indicate that the proposed E_HGSO outperformed the comparison algorithms. Furthermore, we applied E_HGSO to the feature selection problem. The results indicate that E_HGSO performs competitively across various metrics, including, classification accuracy, and