Electrical Engineering and Systems Science > Systems and Control
[Submitted on 3 Nov 2019 (this version), latest version 4 Feb 2021 (v3)]
Title:Stochastic Optimal Control of HVAC system for Energy-efficient Buildings
View PDFAbstract:This paper aims to develop an agile, adaptive and energy-efficient method for HVAC control via Markov decision process (MDP). Our main contributions are outlined First, we formulate the problem as a MDP, which incorporates i) the multiple uncertainties resulting from the weather and occupancy, ii) the elaborate Predicted Mean Vote (PMV) thermal comfort model. Second, to cope with the computational challenges, we propose a gradient-based policy iteration (GBPI) method to learn the policies based on the performance gradients. Thrid, we theoretically prove that the method can converge to an optimal policy of the formulated MDP. The advantages of the proposed method are that: i) it uses off-line computation to learn control policies thus reducing on-line computation burden, and ii) it handles the non-convex and nonlinear system dynamics efficiently and can accommodate the non-analytical thermal comfort models (e.g., PMV) in the literature. The favorable performance of the policies yield by the GBPI is demonstrated through comparisons with the optimal solution obtained by assuming all the information is available before the planning in several case studies.
Submission history
From: Yu Yang [view email][v1] Sun, 3 Nov 2019 06:17:18 UTC (337 KB)
[v2] Fri, 17 Jul 2020 06:24:04 UTC (336 KB)
[v3] Thu, 4 Feb 2021 05:05:19 UTC (337 KB)
Current browse context:
eess.SY
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.