A multi-strategy enhanced Dung Beetle Optimization for real-world engineering problems and UAV path planning

Abstract

Dung Beetle Optimization (DBO) is a widely recognized meta-heuristic algorithm inspired by swarm intelligence. However, it faces significant limitations in convergence speed and solution accuracy, particularly for complex multimodal optimization problems with multiple peaks. To address these challenges, we propose the Enhanced Dung Beetle Optimization (EDBO) algorithm, integrating four innovative mechanisms: (1) an Optimal Value Search Guidance Strategy, utilizing the global best solution to steer the search and mitigate the risk of local optima entrapment; (2) a Nonlinear Dynamic Adjustment Factor, adaptively balancing exploration and exploitation to enhance search diversity across optimization stages; (3) a Preferential Boundary Control Strategy, dynamically refining boundary behavior to direct individuals towards promising regions without stagnation; and (4) an Improved Foraging Enhancement Strategy, incorporating adaptive updates to improve global search efficiency and prevent premature convergence. EDBO was tested on 52 benchmark functions, including CEC 2017, CEC 2020, and CEC 2022, and compared with algorithms like GSA, WOA, LSHADE, and QHDBO. Results show EDBO outperforms these algorithms in convergence speed, accuracy, and stability. Additionally, EDBO was validated on 19 real-world engineering problems and a UAV path planning task, demonstrating its robust global search capabilities and practical applicability. Matlab codes of EDBO are available at https://ww2.mathworks.cn/matlabcentral/fileexchange/179084-a-multi-strategy-enhanced-dung-beetle-optimization.

Keywords

• Dung beetle optimization
• Complex multimodal optimization problems
• Improved search mechanisms
• Boundary control
• Real-world engineering problems
• UAV path planning