Adaptive Multi-Swarm Differential Evolution Algorithm for UAV Path Planning

Path planning for unmanned aerial vehicles (UAVs) remains a crucial prerequisite for UAV application in various fields. However, due to the complexity of the model, several state-of-the-art methods may encounter challenges in finding feasible solutions or be prone to getting stuck in local optima, especially in the complex 3D battlefield environment. An adaptive multi-swarm differential evolution algorithm (AMSDE) is put forward to address these problems. First, we employ the ε-level comparison to ensure the feasibility of the solution. Second, we design an adaptive swarm partitioning technique to avoid crossover evolution between sub-swarms caused by random partitioning. Third, the hierarchical update mechanism is implemented to guide each sub-swarm's search. It facilitates effective communication between sub-swarms and keeps the balance of exploration and exploitation. Experiment results have shown that AMSDE is competitive compared with other excellent algorithms, proving its capability to generate higher-quality paths for UAVs.