A Learning-Based Evolutionary Algorithm for Multi-UAV Cooperative Mission Planning

In this paper, we deal with the cooperative mission planning problem for unmanned aerial vehicles (UAVs) via a learning-based evolutionary optimization algorithm, in which task allocation and trajectory planning are incorporated into a unified framework. Most existing studies handle the task allocation and the trajectory planning separately, thus resulting in poor solutions. Therefore, we propose an integrated graph model of task allocation and trajectory planning optimization, in which the cooperative mission planning problem is transformed to the integer programming problem. Notably, a learning mechanism is developed to discover rules and create new individuals based on the performance of solutions, thereby preventing premature convergence effectively. Simulation results demonstrate the effectiveness and superiority of the proposed algorithm.