Using process data to generate an optimal control policy via apprenticeship and reinforcement learning
AuthorsMowbray, Max; orcid: 0000-0003-1398-0469; email: firstname.lastname@example.org
Smith, Robin; email: email@example.com
Del Rio‐Chanona, Ehecatl A.; email: firstname.lastname@example.org
Zhang, Dongda; orcid: 0000-0001-5956-4618; email: email@example.com
MetadataShow full item record
AbstractAbstract: Reinforcement learning (RL) is a data‐driven approach to synthesizing an optimal control policy. A barrier to wide implementation of RL‐based controllers is its data‐hungry nature during online training and its inability to extract useful information from human operator and historical process operation data. Here, we present a two‐step framework to resolve this challenge. First, we employ apprenticeship learning via inverse RL to analyze historical process data for synchronous identification of a reward function and parameterization of the control policy. This is conducted offline. Second, the parameterization is improved online efficiently under the ongoing process via RL within only a few iterations. Significant advantages of this framework include to allow for the hot‐start of RL algorithms for process optimal control, and robust abstraction of existing controllers and control knowledge from data. The framework is demonstrated on three case studies, showing its potential for chemical process control.
CitationAIChE Journal, page e17306
PublisherJohn Wiley & Sons, Inc.
DescriptionFrom Wiley via Jisc Publications Router
History: received 2020-10-04, rev-recd 2021-04-23, accepted 2021-05-03, pub-electronic 2021-05-15
Article version: VoR
Publication status: Published
Showing items related by title, author, creator and subject.
Identifying barriers to the adoption of Certificated and Experiential Accreditation/Recognition of Prior Learning: A global perspectiveTalbot, Jon; University of Chester (2019-01-25)The presentation reviews research into practice in the UK and beyond to identify barriers to adoption and examples where there has been a systemic increase. The various terms used to describe practice are outlined and its application in the UK and beyond briefly reviewed. The presentation will refer to national, institutional and pedagogical constraints to the wider adoption of practice. Two national examples are cited where practice appears most widespread- the USA and France. Possible explanations are cited and examples of institutional practice in each country described. Finally lessons from a global perspective are used to highlight opportunities and constraints in the UK.
Repurposing MOOCs for the Accreditation of Prior Learning: A survey of practice in university Work Based Learning departmentsTalbot, Jon; University of Chester (Universities Association for Lifelong Learning, 2016-03-18)The presentation summarises a small survey of APL practices in work based learning departments in universities in England and Wales in respect of willingness to accept completion of a MOOC learning programme. The study found few students with MOOC certificates approached universities for accreditation and that few were likely to accept them in any case. The study highlights how many students are now engaged in work based learning and the varieties of practice associated with the Accreditation/ Recognition of Prior Learning.
Who will accredit MOOC learning? a survey of work based learning departments in English and Welsh universitiesTalbot, Jon; University of Chester (The Open University, 2016-02-26)The presentation is of a small survey to determine whether Work based learning departments have sufficient flexibility to admit MOOC certificates as the basis for APL/RPL claims. The main finding is that there is low awareness of MOOCs among tutors such that is unlikely many would recognise the value of a MOOC certificate as the basis for a claim for past learning.