基于自适应探索改进的深度增强学习算法

毛坚桓; 殷璐嘉

Chinese

您当前的位置：

首页 >

文章列表页 >

基于自适应探索改进的深度增强学习算法

更新时间：2026-09-04

- 基于自适应探索改进的深度增强学习算法
- A Deep Reinforcement Learning Algorithm Based on Adaptive Exploration
- 微电子学与计算机 2016年33卷第6期页码：139-142
- 作者机构：
  
  中国电子科技集团公司第二十八研究所,江苏,南京,210007
- 作者简介：
- 基金信息：
- DOI：
  中图分类号：
- 纸质出版：2016
- 稿件说明：
移动端阅览
毛坚桓, 殷璐嘉. 基于自适应探索改进的深度增强学习算法[J]. 微电子学与计算机, 2016, 33(6): 139-142.

MAO Jian-huan, YIN Lu-jia. A Deep Reinforcement Learning Algorithm Based on Adaptive Exploration[J]. 2016, 33(6): 139-142.
毛坚桓, 殷璐嘉. 基于自适应探索改进的深度增强学习算法[J]. 微电子学与计算机, 2016, 33(6): 139-142. DOI：

MAO Jian-huan, YIN Lu-jia. A Deep Reinforcement Learning Algorithm Based on Adaptive Exploration[J]. 2016, 33(6): 139-142. DOI：

摘要

针对深度增强学习算法中探索开发策略的平衡问题

提出一种基于VDBE(Value-Difference Based Exploration)扩展的自适应探索改进算法.该算法依赖于值函数差异提出一种基于状态的探索控制策略

以达到在初始学习阶段不熟悉周围环境时agent采取积极探索策略

而随着深入学习和周边环境的熟悉

agent逐渐降低探索率的自适应探索/开发平衡的理想行为状态.

Abstract

To find a balance between exploration and exploitation

this paper proposes a VDBE(Value-Difference Based Exploration) based algorithm. The algorithm proposes a state-based control strategy depends on the value difference. In order to achieve the ideal exploration/exploitation behavior state

agent takes positive actions to explore environments in the initial stage of learning when agent is unfamiliar with surrounding environment. As learning time goes on and agent is more familiar with surrounding

it gradually reduces the exploration rate.

关键词

Keywords

references

Long-Ji Lin. Reinforcement learning for robots using neural networks[R]. Technical report, DTIC Document, 1993.

Sutton RS, Barto AG. Reinforcement Learning: An Introduction[M]. MIT Press, Cambridge, MA: 1998.

Watkins C. Learning from delayed rewards[D]. Cambridge: University of Cambridge, 1989.

Bertsekas D P. Dynamic Programming: Deterministic and Stochastic Models[M]. Englewood Cliffs, NJ: Prentice-Hall, 1987.

ToddHester, Manuel Lopes, Peter Stone. Learning exploration strategies in model-based reinforcement learning[C]// The Twelfth International Conference on Autonomous Agents and Multiagent Systems (AAMAS), St Paul, Minnesota: IEEE, 2013.

Brafman R I, Tennenholtz M. R-MAX a general polynomial time algorithm for near-optimal reinforcement learning[R]. Journal of Machine Learning Research, 2002, 21(3): 213-231.

Thrun S B. Efficient exploration in reinforcement learning[R]. Technical Report CMU-CS-92-102, Carnegie Mellon University, Pittsburgh, PA, USA: 1992.

Ishii S, Yoshida W, Yoshimoto J. Control of exploitation-exploration metaparameter in reinforcement learning[J]. Neural Networks, 2002, 15(4-6): 665-687. DOI: 665

Volodymyr Mnih, Koray Kavukcuoglu, David Silver, et al. Playing Atari with Deep Reinforcement Learning[R]. arXiv preprint arXiv: 2013.

Kaelbling L P, Littman M L, Moore A W. Reinforcement Learing: a survey[J]. Journal of Artificial Intelligence Research, 1996, 22(4): 237-285. DOI: 237

Michel Tokic. Adaptive ε-greedy exploration in reinforcement learning based on value differences[C]// Advances in Artificial Intelligence 33rd Annual German Conference on AI, Karlsruhe, Germany, IEEE, 2010: 21-24.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于边缘云计算的5G通信信道低延迟传输自适应控制方法

质心不重合的移动机器人自适应滑模轨迹跟踪控制

V型沟槽SiC中子探测器的蒙特卡洛分析

低功耗设计的双链电源关断策略的实现与验证

超低功耗实时时钟电路的设计