基于启发式强化学习的多智能体覆盖问题研究

Multi-Agent Coverage Based on Heuristically Accelerated Reinforcement Learning

针对多智能体覆盖问题存在的计算量大、收敛速度慢等问题,提出一种基于启发式强化学习的多智能体覆盖算法。利用智能体收集到的环境信息作为先验知识,对强化学习中智能体的行动选择进行引导。仿真实验表明,该算法在不影响覆盖效果的情况下有效提高覆盖问题的学习收敛速度。

Since multi-agent coverage problem requires large amount of computation and can be very time consuming, proposes a multi-agent coverage algorithm based on heuristically accelerated reinforcement learning. Agents extract the environmental structure information as priori knowledge to guide the choices of actions in reinforcement learning. The simulation results show that the algorithm can effectively speed up the learning convergence of the coverage problem without affect the coverage result.