{"title":"Reinforcement learning for optimal control of stochastic nonlinear systems","authors":"Xinji Zhu, Yujia Wang, Zhe Wu","doi":"10.1002/aic.18840","DOIUrl":null,"url":null,"abstract":"A reinforcement learning (RL) approach is developed in this work for nonlinear systems under stochastic uncertainty. A stochastic control Lyapunov function (SCLF) candidate is first constructed using neural networks (NNs) as an approximator to the value function, and then a control policy designed using this SCLF is developed to ensure the stability in probability of the stochastic nonlinear system. An RL algorithm is proposed for stochastic nonlinear systems to iteratively update the value function and control policy, driving them toward optimal values. We further extend its feasibility under the sample-and-hold implementation of control actions, and demonstrate its application to two chemical reactor examples to show its practical advantages and efficiency.","PeriodicalId":120,"journal":{"name":"AIChE Journal","volume":"72 1","pages":""},"PeriodicalIF":3.5000,"publicationDate":"2025-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"AIChE Journal","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1002/aic.18840","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENGINEERING, CHEMICAL","Score":null,"Total":0}
引用次数: 0
Abstract
A reinforcement learning (RL) approach is developed in this work for nonlinear systems under stochastic uncertainty. A stochastic control Lyapunov function (SCLF) candidate is first constructed using neural networks (NNs) as an approximator to the value function, and then a control policy designed using this SCLF is developed to ensure the stability in probability of the stochastic nonlinear system. An RL algorithm is proposed for stochastic nonlinear systems to iteratively update the value function and control policy, driving them toward optimal values. We further extend its feasibility under the sample-and-hold implementation of control actions, and demonstrate its application to two chemical reactor examples to show its practical advantages and efficiency.
期刊介绍:
The AIChE Journal is the premier research monthly in chemical engineering and related fields. This peer-reviewed and broad-based journal reports on the most important and latest technological advances in core areas of chemical engineering as well as in other relevant engineering disciplines. To keep abreast with the progressive outlook of the profession, the Journal has been expanding the scope of its editorial contents to include such fast developing areas as biotechnology, electrochemical engineering, and environmental engineering.
The AIChE Journal is indeed the global communications vehicle for the world-renowned researchers to exchange top-notch research findings with one another. Subscribing to the AIChE Journal is like having immediate access to nine topical journals in the field.
Articles are categorized according to the following topical areas:
Biomolecular Engineering, Bioengineering, Biochemicals, Biofuels, and Food
Inorganic Materials: Synthesis and Processing
Particle Technology and Fluidization
Process Systems Engineering
Reaction Engineering, Kinetics and Catalysis
Separations: Materials, Devices and Processes
Soft Materials: Synthesis, Processing and Products
Thermodynamics and Molecular-Scale Phenomena
Transport Phenomena and Fluid Mechanics.