Abstract
This study proposes Enhanced Beaver Behavior Optimizer (EBBO) to overcome the original BBO algorithm's limitations in handling complex optimization problems. EBBO integrates a three-phase cooperative framework, incorporating adaptive mutation, dynamic opposition-based learning, and an risk-aware decision strategy inspired by simulated annealing. Comprehensive evaluations on the CEC 2017 and CEC 2020 benchmark suites demonstrate that EBBO significantly outperforms nine widely used algorithms (e.g., BBO, FATA, DE) in convergence accuracy, stability, and robustness, especially for high-dimensional and multimodal functions. EBBO achieves average objective value reductions of 15-50% and standard deviation reductions of 30-70% compared to the original BBO, with Wilcoxon rank-sum tests confirming statistical significance across most functions. When applied to three classical engineering design problems-step-cone pulley, pressure vessel, three-bar truss optimization, and 3D UAV path planning-EBBO consistently achieved the best or near-optimal solutions while satisfying all nonlinear constraints. The results confirm that EBBO effectively balances exploration and exploitation, offering a reliable and efficient approach for solving complex constrained optimization challenges in both benchmark and real-world engineering contexts.