An improved prairie dog optimization algorithm integrating multiple strategies and its application

Abstract

Aiming at the problems in prairie dog optimization (PDO), such as uneven population distribution at initialization, slow convergence, imbalance between global exploration and local exploitation, and the tendency to get trapped in the local optimum, this study proposes an Improved prairie dog optimisation algorithm integrating multiple strategies (IMSPDO). Firstly, the population is initialized using spatial pyramid matching (SPM) chaotic mapping combined with improved random opposition-based learning (IROL) to solve the problems of uneven distribution and poor diversity of the population. Secondly, the prey escapes energy formula mentioned in the harris hawks optimization (HHO) is introduced to achieve the smooth transition between the exploration phase and the exploitation phase, balancing the algorithm’s global exploration capability and local exploitation capability. Additionally, the idea of the particle swarm optimization (PSO) is applied to enhance the global optimization capability of the algorithm. Finally, the ideas of simulated annealing (SA), polynomial mutation and Cauchy mutation are also introduced to improve the ability that individuals to jump out of the local optimum. The performance of the improved algorithm is verified on a set of 21 classical benchmark functions and 8 CEC2020 test functions. The proposed IMSPDO is also evaluated against original PDO, and six other commonly used algorithms. The result of the Wilcoxon rank-sum test shows that there is a significant difference between the selected algorithms and IMSPDO. Furthermore, 3 engineering examples are used to further test the superiority of IMSPDO in dealing with real-world problems.