A hybrid evolutionary algorithm integrated with specific neighborhood search for collaborative task planning of heterogeneous multi-UAVs under SEAD scenarios

To overcome the challenges posed by diversified task combinations and strict temporal constraints in the Suppression of Enemy Air Defenses (SEAD), this paper investigates the task planning problem. The task correlation in SEAD evolves from chained correlations within individual units to a web of interdependencies across all UAVs, making the task planning problem significantly more complex than traditional vehicle routing problems. This complexity ensures that any alteration to a single task will produce a cascading effect on the overall plan. To address this issue, the paper designs a matrix-based task dependency analysis method that quantitatively evaluates the logical interdependencies between tasks. Additionally, the concept of dynamic time windows is introduced to precisely measure the temporal impact of task adjustments on the overall plan. On the basis of these analyses, an adaptive hybrid evolutionary algorithm is designed, which employs a neighborhood search strategy to precisely adjust task assignments and execution sequences based on task dependencies and dynamic time windows. Finally, simulation results suggest that the algorithm proposed in this paper consistently produces task plans with a high number of completed tasks and short completion time across scenarios involving varying task quantities, task combinations, time windows, regional distributions, and UAV configurations.