基于粒子滤波的强化学习算法研究

Research on reinforcement learning algorithm based on particle filter

文章分析了一种基于粒子滤波和强化学习的算法。该算法通过结合粒子滤波和Q-学习算法,得到一种基于粒子滤波和强化学习的机会频谱接入算法(RLPF)。实验结果表明,RLPF算法能够在策略空间直接进行全局搜索,这是对传统的基于局部搜索策略的强化学习算法的明显改善。

This paper analyzes a particle filter based on reinforcement learning algorithm. The new algorithm processed the opportunistic spectrum access algorithm based on particle filter and Q-learning algorithm by combining particle filter and Q-learning algorithm. The experimental results show that the RLPF algorithm can be directly used for global search in the strategy space, which is a significant improvement of traditional reinforcement learning algorithm based on the local search strategy.