AR-RBMO: An enhanced red-billed blue magpie optimizer with attraction-repulsion and dynamic balancing strategies for global optimization

Abstract Metaheuristic algorithms have been extensively applied to real-world optimization problems because of their flexibility and strong problem-solving ability. However, as optimization problems become increasingly complex and diverse, stand-alone algorithms encounter inherent limitations that diminish their effectiveness. The red-billed blue magpie optimizer (RBMO), a relatively new swarm intelligence algorithm, has demonstrated significant potential, while its performance remains limited by restricted global exploration capability a tendency to converge prematurely to local optima. Combining the strengths of multiple algorithms enables the creation of more effective hybrid optimization methods. Building on this idea, this study introduces an attraction–repulsion enhanced red-billed blue magpie optimizer (AR-RBMO). The algorithm incorporates an attraction–repulsion mechanism to improve global search, a best-solution attraction strategy to direct the population towards high-quality regions, an escape strategy to avoid local optima, and a dynamic exploration–exploitation balance strategy based on optimal solution feedback. Systematic experiments on the CEC2017 benchmark suite, covering 30-, 50-, and 100-dimensional functions, evaluate AR-RBMO against 18 representative metaheuristic algorithms. Results from the Friedman test indicate average rankings of 2.133, 1.75, and 1.4667, respectively, confirming AR-RBMO’s overall superiority. The Wilcoxon rank test further validates that these performance improvements are statistically significant. Evaluation on six classical engineering optimization problems yields high-quality solutions, demonstrating robust global search capabilities, high convergence accuracy, and consistent solution stability.