Multi-UAV Air Combat Weapon-Target Assignment Based On Genetic Algorithm And Deep Learning

Considering the actual process of multiple unmanned aerial vehicles(UAV) air combat, this paper mainly studies how to intercept the enemy's multiple UAVs in the stage of beyond visual range (BVR). As for the problem of weapon launch beyond line-of-sight, it is essentially the problem of weapon-target assignment(WTA). Taking into account the random search capabilities based on genetic algorithms and the powerful feature extraction and learning capabilities of deep neural networks, a more efficient solution is proposed. Through two typical scenarios of algorithm experiments, it is verified that GA-DL model can not only learn the decision-making idea of genetic algorithm, but also greatly improve the running speed of the algorithm. It provides a new idea for realizing the automatic process of weapon-target assignment problem solving.