Attention-Embedded Lion Pride Labor Division Algorithm for Solving Dynamic Coupling SEAD Task Allocation Problem

As an evolving frontier in the application of unmanned aerial vehicles (UAVs), the incorporation of labor division intelligence from natural biological groups into UAV swarms is garnering heightened attention and implementation. This approach is distinguished by the execution of complex tasks via multi-UAV collaboration, along with dynamic adaptation to the environment through interactions with targets and peers, all without human oversight. With this objective in mind, this paper delves into a specific set of problems related to the task allocation for the suppression of enemy air defenses (DC-SEAD) within dynamic and interrelated environments, a scenario prevalent in real-world combat situations. We introduce an optimizer, the attention-embedded lion pride labor division algorithm (A-LPLDA). It begins by constructing a task allocation model for UAV swarms executing the DC-SEAD mission, taking into account various constraints and dynamic operational factors. Following this, we establish a self-organized optimizer inspired by the labor division behaviors observed in lion prides. This optimizer independently formulates solutions through individual interactions, disconnected from the objective function, thereby ensuring remarkable adaptability and scalability. Furthermore, we integrate attention mechanisms into the UAV swarm task allocation process. This includes the development of enhancement mechanisms for several UAV interference scenarios, aimed at augmenting task execution efficiency. To evaluate the algorithm's effectiveness, we have crafted multiple representative scenario sets that consider a variety of dynamic factors, the overall influence of these factors within the environment, and the quantity of targets and UAVs involved. The experimental outcomes affirm the superior performance of the proposed A-LPLDA.