ABSTRACT A multi‐strategy improved dung beetle optimiser (MSDBO) is proposed for unmanned aerial vehicle (UAV) 3D path planning, aiming to solve the limitations of traditional algorithms (A*, Dijkstra's) in dynamic and multi‐constraint 3D spaces. The MSDBO optimises the original dung beetle optimiser (DBO) by three strategies: Fuch map initialisation for better population diversity, opposition‐based learning for faster convergence, and adaptive inertia weight for balancing global exploration and local exploitation. These improvements effectively enhance the algorithm's convergence speed, solution quality, and anti‐local optimum ability. MATLAB simulation experiments show that MSDBO outperforms other mainstream optimisation algorithms in fitness value and path length for UAV 3D path planning.
Lin et al. (Thu,) studied this question.