- 1、本文档共33页,可阅读全部内容。
- 2、有哪些信誉好的足球投注网站(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
- 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
Mach Learn
DOI 10.1007/s10994-011-5235-x
Reinforcement learning in feedback control
Challenges and benchmarks from technical process control
Roland Hafner · Martin Riedmiller
Received: 26 February 2010 / Revised: 3 January 2011 / Accepted: 8 January 2011
© The Author(s) 2011
Abstract Technical process control is a highly interesting area of application serving a high
practical impact. Since classical controller design is, in general, a demanding job, this area
constitutes a highly attractive domain for the application of learning approaches—in par-
ticular, reinforcement learning (RL) methods. RL provides concepts for learning controllers
that, by cleverly exploiting information from interactions with the process, can acquire high-
quality control behaviour from scratch.
This article focuses on the presentation of four typical benchmark problems whilst high-
lighting important and challenging aspects of technical process control: nonlinear dynamics;
varying set-points; long-term dynamic effects; influence of external variables; and the pri-
macy of precision. We propose performance measures for controller quality that apply both
to classical control design and learning controllers, measuring precision, speed, and stability
of the controller. A second set of key-figures describes the performance from the perspec-
tive of a learning approach while providing information about the efficiency of the method
with respect to the learning effort needed. For all four benchmark problems, extensive and
detailed information is provided with which to carry out the evaluations outlined in this
article.
A close evaluation of our own RL learning scheme, NFQCA (Neural Fitted Q Iteration
with Continuous Actions), in acordance with the proposed scheme on all four benchmarks,
thereby provides performance figures on both control quality and learning behavior.
Keywords Reinforcement learning · Feedback control · Benchmarks · Nonlinear control
1 I
您可能关注的文档
- 垂直绿化--拓宽城市绿化空间的有效途径.pdf
- 可视化管理实施方案.pdf
- 通信机房标准化建设研究.pdf
- 板翅式换热器参数化绘图.pdf
- 小区智能化系统详细说明.pdf
- GML数据的Web可视化设计与实现.pdf
- PACS与医学摄影图像数字化管理.pdf
- 海河水质咸化原因的初步分析.pdf
- 中国特色城镇化道路研究.pdf
- 儿童家具人性化设计.pdf
- 中国国家标准 GB/T 18233.4-2024信息技术 用户建筑群通用布缆 第4部分:住宅.pdf
- GB/T 18233.4-2024信息技术 用户建筑群通用布缆 第4部分:住宅.pdf
- GB/T 18978.210-2024人-系统交互工效学 第210部分:以人为中心的交互系统设计.pdf
- 《GB/T 18978.210-2024人-系统交互工效学 第210部分:以人为中心的交互系统设计》.pdf
- 中国国家标准 GB/T 18978.210-2024人-系统交互工效学 第210部分:以人为中心的交互系统设计.pdf
- GB/T 16649.2-2024识别卡 集成电路卡 第2部分:带触点的卡 触点的尺寸和位置.pdf
- 《GB/T 16649.2-2024识别卡 集成电路卡 第2部分:带触点的卡 触点的尺寸和位置》.pdf
- 中国国家标准 GB/T 16649.2-2024识别卡 集成电路卡 第2部分:带触点的卡 触点的尺寸和位置.pdf
- GB/T 17889.4-2024梯子 第4部分:铰链梯.pdf
- 《GB/T 17889.4-2024梯子 第4部分:铰链梯》.pdf
文档评论(0)