Abstract The equilibrium optimizer (EO) is a novel physics-based metaheuristic approach that draws inspiration from the dynamic mass balance in physics. While it has exhibited promising results in addressing numerical and engineering design problems, it has several limitations, including low diversification, poor exploration ability, and stagnation in local optima. To surmount these limitations, this research addresses an enhanced EO (AEO) method that integrates strategies to establish a more harmonized balance between exploration and exploitation. These mechanisms are the adaptive elite-guided search mechanism and interparticle information interaction strategy. Each mechanism fulfills a distinct role in the search process. The adaptive elite-guided search focuses on improving exploitation capability and evading local optima. Meanwhile, interparticle information interaction facilitates the promotion of population diversity. The synergistic interplay between these dual strategies serves to refine the balance between exploitation and exploration. A comprehensive series of experiments is conducted to investigate the efficiency of the reported algorithm. The results of AEO are compared with a wide range of metaheuristic techniques, including the basic EO, well-known EO variants, and recently reported advanced metaheuristics. Experimental findings indicate that AEO consistently surpasses comparison optimization algorithms in 77.78% of the benchmark tests, while also delivering exceptional results in 95.65% of the high-dimensional benchmarks. Quantitative and qualitative analysis results demonstrate the superiority and robustness of the developed algorithm compared to its competitors. The statistical robustness of the performance is also confirmed through the utilization of the Wilcoxon signed-rank test. Furthermore, the applicability of AEO is investigated by implementing it as a mobile robot path-planning technique. Comparative assessments against well-known metaheuristics illustrate the favorable potential of the proposed algorithms as promising path planners.
Wang et al. (Wed,) studied this question.