Efficient multi-UAV path planning in dynamic and complex environments using hybrid polar lights optimization
Zhi Xu, Zhilin Wang, Rui Liu, Chenliang Huang, Yuxiang Shi, Mingjing Wang, Huiling Chen · Journal of King Saud University - Computer and Information Sciences · 2025
Unmanned Aerial Vehicle (UAV) swarm path planning poses significant challenges, particularly in dynamic environments with complex obstacles. Traditional path-planning methods often encounter difficulties related to high dimensionality and obstacle density. This paper introduces a novel MA based on artificial intelligence, termed Improved Polar Lights Optimization (CCPLO). The CCPLO enhances path planning performance by integrating the Criss-Cross (CC) strategy with the Covariance Matrix Adaptation Evolution Strategy (CMA-ES). Key improvements include a dynamic adjustment mechanism for search parameters, the implementation of parallel processing techniques, and refined crossover and normalization processes. The dynamic adjustment mechanism allows for real-time tuning of parameters, enhancing global search capabilities and minimizing the risk of local optima. Furthermore, parallel processing significantly boosts computational efficiency, especially in high-dimensional and complex scenarios. Experimental results demonstrate that the CCPLO algorithm outperforms existing algorithms in the CEC 2017 benchmark function test set. Specifically, in multi-obstacle and dynamic task environments, CCPLO effectively designs safer and more efficient paths, highlighting its strong potential for UAV swarm path planning.