Abstract
Cooperative path planning of multiple unmanned aerial vehicles (UAVs) is pivotal for improving mission efficiency and safety in complex scenarios. However, the multi-constraint of UAVs increases the design difficulity of cooperative path planning. To address these issues, a hybrid search behavior-based adaptive grey wolf optimizer (HSB-GWO) is proposed in this work. HSB-GWO incorporates three key innovations: (1) A dimension learning-based hunting (DLH) strategy is employed to enhance population diversity by enabling knowledge exchange between non-leader wolves and their neighbors. (2) Aquila exploration combining expand exploration for global potential region detection and Lévy flight-based narrowed exploration for preventing populations from falling into local optimal solutions is adopted to enrich search behaviors and avoid local optima. (3) An adaptive weight adjustment mechanism is designed for leader wolves (α, β, and δ) to dynamically tune their contribution to offspring generation based on fitness to improve high-quality solution utilization. The search performance of HSB-GWO on the benchmark functions was validated by experiments on the benchmark suites of IEEE CEC 2017 and 2019, in which HSB-GWO outperformed seven comparison algorithms (AO, AOA, CBOA, NOA, GWO, IGWO, and AGWO), with Friedman test confirming its top overall rank (Rank 1). The results of cooperative path planning simulation demonstrate that the high-quality multi-UAV trajectories can be generated by the HSB-GWO to guide UAVs from the start to the destination safely and smoothly with the smallest cost.