Multiple Targets Assignment Based on Reinforcement Learning Hyper-Heuristic Algorithm of Multi-UAV

Aiming at the requirements for high speed and high accuracy in cooperative task assignment for multi-UAV, a reinforcement learning based hyper-heuristic multi-task assignment algorithm is developed in this paper. Taking into account task priority, flight cost, and task risk, a object function is constructed for multi-UAV cooperative task assignment. Four lowlevel heuristic (LLH) operators are designed, and combined with the reward-punishment mechanism of Q-learning and the concept of hyper-heuristics, a high-level selection strategy is obtained based on reinforcement learning. The operator is selected by evaluating the performance of the LLH operators. The simulation results demonstrate that the reinforcement learning based hyperheuristic multi-task assignment algorithm can accurately and effectively address the multi-UAV cooperative task assignment problem.