Moonlight Bat Optimization (MBO): A Nature-inspired Metaheuristic Balancing Adaptive Exploration and Precision Exploitation

Abstract

Moonlight Bat Optimization (MBO) algorithm, a novel bio-inspired metaheuristic that emulates the nocturnal foraging behavior of Moonlight Bats, is proposed in this paper. MBO integrates two complementary phases—global exploration and local exploitation—to achieve a robust balance between search diversity and convergence precision. The exploration phase is inspired by high-altitude, wide-area flights, where stochastic, moonlight-scaled movements and frequency-modulated attraction toward promising regions prevent premature convergence and promote comprehensive coverage of the solution space. The exploitation phase mimics low-altitude precision hunting, applying directed, distance-aware adjustments and loudness-scaled local perturbations to refine candidate solutions near high-fitness areas. The algorithm was rigorously evaluated on 23 benchmark functions, encompassing unimodal, high-dimensional multimodal, and fixed-dimensional multimodal problems, and compared against nine advanced metaheuristic algorithms. Results demonstrate that MBO consistently achieves competitive convergence rates and high-quality solutions, effectively preserving population diversity while exploiting promising regions. However, it is noteworthy that MBO does not attain the top rank on all test functions, highlighting inherent challenges in complex multimodal landscapes and indicating potential avenues for algorithmic enhancement. Key contributions of this work include: the biologically grounded formulation of exploration and exploitation operators, rigorous mathematical modeling of bat-inspired search behaviors, and comprehensive comparative performance analysis. MBO provides a flexible and interpretable framework suitable for diverse complex optimization tasks, and its design principles offer opportunities for extensions to constrained, dynamic, and large-scale optimization problems.

Keywords: Adaptive algorithm, Computational intelligence, Global exploration, Local exploitation, Metaheuristic, Moonlight bat optimization, Multimodal optimization