Dynamic multi-strategy Grey Wolf optimizer and its applications

To address the shortcomings of the Grey Wolf Optimizer (GWO) in solution accuracy, convergence speed, and search capability, this paper proposes a Dynamic Multi-Strategy Grey Wolf Optimizer (DMSGWO). Based on GWO, DMSGWO introduces four improvement strategies. First, a nonlinear convergence factor strategy is adopted to better balance global exploration and local exploitation. Second, a population dynamic grouping strategy is employed to dynamically adjust the population sizes of exploration and exploitation groups during the iteration process. Third, a random position update strategy is applied to the exploration group to enhance the global exploration capability. Finally, an adaptive perturbation position update strategy is applied to the exploitation group to improve the local exploitation capability. To comprehensively evaluate the performance of DMSGWO, it is compared with GWO, four other swarm intelligence optimization algorithms, and four other improved GWO algorithms on 23 benchmark test functions and the CEC2022 test suite. The comparative results demonstrate that DMSGWO exhibits excellent solution accuracy, stability, and convergence speed. Further validation through Friedman ranking tests and Wilcoxon signed-rank tests confirms that DMSGWO ranks first in overall performance. The test results show statistically significant performance differences from the other algorithms on most test functions, and further verify the superior performance of DMSGWO. To validate the practical application value of DMSGWO, it is applied to two typical engineering design optimization problems and WSN coverage optimization problem alongside the comparative algorithms. The results show that DMSGWO can effectively handle constraints and achieve the best objective values. Thus, these results demonstrates the effectiveness and practicality of DMSGWO in solving real-world complex application problems.