论文部分内容阅读
为了解决模糊Sarsa学习(FSL)无法在线自适应调节学习因子和不能处理学习过程中探索与利用的平衡问题,提出了一种改进的模糊Sarsa学习(IFSL)算法.在FSL基础上,引入自适应学习率产生器来在线调节学习因子,增加模糊平衡器控制探索和利用的程度.给出了IFSL的结构框图,证明了IFSL中可调节权向量具有平衡不动点.仿真结果表明,与FSL相比,IFSL能加快系统的学习收敛速度,具有较好的学习性能.
In order to solve the problem that fuzzy Sarsa learning (FSL) can not adaptively adjust learning factors online and can not deal with the balance of exploration and utilization in learning process, an improved fuzzy Sarsa learning (IFSL) algorithm is proposed. Based on FSL, Learning rate generator to adjust the learning factor online and increase the degree of exploration and utilization of the fuzzy balancer control.The structural block diagram of IFSL is given and it is proved that the adjustable weight vector in IFSL has a balanced fixed point.The simulation results show that, Compared to, IFSL can speed up the learning convergence speed of the system, and has good learning performance.