| 一类基于有效跟踪的广义平均奖赏激励学习算法 |
| |
| 引用本文： | 陈焕文,谢建平. 一类基于有效跟踪的广义平均奖赏激励学习算法[J]. 计算机工程与应用, 2002, 38(1): 65-68 |
| |
| 作者姓名： | 陈焕文 谢建平 |
| |
| 作者单位： | 1. 长沙电力学院数学与计算机系,长沙,410077
2. 长沙交通学院网络中心,长沙,410076 |
| |
| 基金项目： | 国家自然科学基金，湖南省教育厅科研基金 |
| |
| 摘 要： | 取消了平均奖赏激励学习的单链或互通MDPs假设,基于有效跟踪技术和折扣奖赏型SARSA(λ)算法,对传统的平均奖赏激励学习进行了推广,提出了一类广义平均奖赏激励学习算法,并对算法的性能进行了初步的比较实验。
|
| 关 键 词： | 激励学习 Markov决策过程 平均奖赏 有效跟踪 |
| 文章编号： | 1002-8331-(2002)01-0065-04 |
| 修稿时间： | 2001-06-01 |
A Class of Generalized Algorithms of Average-Reward Reinforcement Learning Based on Eligibility Traces |
| |
| Chen Huanwen Xie Jianping. A Class of Generalized Algorithms of Average-Reward Reinforcement Learning Based on Eligibility Traces[J]. Computer Engineering and Applications, 2002, 38(1): 65-68 |
| |
| Authors: | Chen Huanwen Xie Jianping |
| |
| Affiliation: | Chen Huanwen 1 Xie Jianping 21 |
| |
| Abstract: | The assumption of unichain or communicating MDPs in average-reward reinforcement learning has been taken off.The classical methods of average -reward reinforcement learning are generalized with eligibility traces and discounted SARSA(λ)algorithm.A class of generalized algorithms for the average -reward reinforcement learning is proposed,and preliminary empirical results are presented to compare the performance of these new algorithms. |
| |
| Keywords: | Reinforcement learning Markov decision processes(MDPs ) Average rewards Eligibility traces. |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
