Abstract
A new enhanced grasshopper optimization algorithm (GOA) has been developed and successfully applied to feature selection. GOA, as a heuristic algorithm, is proposed by simulating the living habits of grasshoppers in nature. Although GOA has an excellent global optimization capability, it still faces the disadvantage of low efficiency of searching optimization due to its ease of falling into the local optimum. Hence, based on the original GOA, this study integrates new ideas to reduce the defects to obtain a better global optimization ability. Because of the continuous optimization problem, the features of pursuing the best possible individual of spiral motion have been considered. The spiral motion is integrated into the GOA exploitation search stage, which further expands the diversification and intensification trends' capacities and effectively balances the exploration and exploitation procedures. Intuitively speaking, GOA with spiral search method can find better solutions in the exploration movement process, which is more efficient than the original search method. In the experimental comparison, to verify the proposed SGOA's ability in dealing with global unconstrained and constrained optimization problems, we compared it with other 30 IEEE 2017 benchmark tasks in meta-heuristic algorithms. Then, it is adopted to optimize engineering design and feature selection problems. We can know that the proposed SGOA has a good optimization ability in practical application from the experimental results. Spiral motion mode can significantly improve the original GOA's exploitation and exploration ability, and the proposed SGOA is of great assistance in practical fields. More info about this paper can be found on the web services https://aliasgharheidari.com .