题名 | Evolving Constrained Reinforcement Learning Policy |
作者 | |
通讯作者 | Jialin Liu |
DOI | |
发表日期 | 2023
|
会议名称 | International Joint Conference on Neural Networks (IJCNN)
|
ISSN | 2161-4393
|
ISBN | 978-1-6654-8868-6
|
会议录名称 | |
页码 | 1-8
|
会议日期 | 18-23 June 2023
|
会议地点 | Gold Coast, Australia
|
出版地 | 345 E 47TH ST, NEW YORK, NY 10017 USA
|
出版者 | |
摘要 | Evolutionary algorithms have been used to evolve a population of actors to generate diverse experiences for training reinforcement learning agents, which helps to tackle the temporal credit assignment problem and improves the exploration efficiency. However, when adapting this approach to address constrained problems, balancing the trade-off between the reward and constraint violation is hard. In this paper, we propose a novel evolutionary constrained reinforcement learning (ECRL) algorithm, which adaptively balances the reward and constraint violation with stochastic ranking, and at the same time, restricts the policy's behaviour by maintaining a set of Lagrange relaxation coefficients with a constraint buffer. Extensive experiments on robotic control benchmarks show that our ECRL achieves outstanding performance compared to state-of-the-art algorithms. Ablation analysis shows the benefits of introducing stochastic ranking and constraint buffer. |
关键词 | |
学校署名 | 第一
; 通讯
|
语种 | 英语
|
相关链接 | [IEEE记录] |
收录类别 | |
资助项目 | National Natural Science Foundation of China[
|
WOS研究方向 | Computer Science
; Engineering
|
WOS类目 | Computer Science, Artificial Intelligence
; Computer Science, Hardware & Architecture
; Engineering, Electrical & Electronic
|
WOS记录号 | WOS:001046198707051
|
来源库 | IEEE
|
全文链接 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10191982 |
引用统计 |
被引频次[WOS]:0
|
成果类型 | 会议论文 |
条目标识符 | http://sustech.caswiz.com/handle/2SGJ60CL/553192 |
专题 | 工学院_斯发基斯可信自主研究院 工学院_计算机科学与工程系 |
作者单位 | 1.Research Institute of Trustworthy Autonomous Systems (RITAS), Southern University of Science and Technology, Shenzhen, China 2.Department of Computer Science and Engineering, Guangdong Key Laboratory of Brain-inspired Intelligent Computation, Southern University of Science and Technology, Shenzhen, China |
第一作者单位 | 斯发基斯可信自主系统研究院 |
通讯作者单位 | 计算机科学与工程系 |
第一作者的第一单位 | 斯发基斯可信自主系统研究院 |
推荐引用方式 GB/T 7714 |
Chengpeng Hu,Jiyuan Pei,Jialin Liu,et al. Evolving Constrained Reinforcement Learning Policy[C]. 345 E 47TH ST, NEW YORK, NY 10017 USA:IEEE,2023:1-8.
|
条目包含的文件 | 条目无相关文件。 |
|
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论