An effective method for global optimization - Improved slime mould algorithm combine multiple strategies

The stochastic search algorithms are an important optimization technique used to solve complex global optimization problems. The Slime Mould Algorithm (SMA) is one of stochastic search algorithm inspired by the observed behaviors and morphological changes in the foraging process of slime moulds. SMA has the advantage of having few parameters and a simple structure, making it applicable to various real-world optimization problems. However, it also has some drawbacks, such as high randomness during the search process and a tendency to converge to local optima, resulting in decreased accuracy. Therefore, we propose an effective method for global optimization - improved slime mould algorithm combine multiple strategy, called EISMA. In EISMA, we introduce a method of exploration that combines the average position of the population with Levy flights to enhance the algorithm's search capability in the previous phase. Then, a novel information-exchange hybrid elite learning operator is proposed to improve the guidance ability of the best search agent. Finally, a dual differential mutation search method that combines global and local optimization is introduced to maintain the diversity of the population by updating the search agents obtained in each iteration. These operations facilitate the algorithm's ability to escape local optima and ensure continuous optimization. To validate the applicability of EISMA, we numerically test it on 39 benchmark functions from CEC2013 and CEC2017 and compare its performance with SMA, as well as 7 modified, 5 standard and 4 classic stochastic search algorithms. Experimental results demonstrate that EISMA outperforms other versions in terms of optimization search performance. Furthermore, EISMA has achieved promising outcomes in testing problems related to path planning for threedimensional unmanned aerial vehicles, pressure vessel design and robot gripper problem.