Author
Listed:
- Ahmed, Bilal
- Ahmed, Ateeque
- Kathirgamanathan, Anjukan
- Guo, Jun-Liang
- Zhang, Hong-Na
- Li, Xiao-Bin
- Qu, Kai-Yang
- Li, Feng-Chen
Abstract
The increasing frequency and severity of climate-driven heatwaves intensifies cooling demand in buildings, exacerbating peak electricity loads and threatening grid stability. Large multizone office buildings pose challenges with complex thermal dynamics, delayed HVAC responses and strict thermal comfort requirements. Reinforcement learning has shown promise for building energy management, but existing studies focus on nominal weather conditions, small scale buildings, dual-objective optimization and instantaneous state representations, limiting their performance under extreme heat events. To address these gaps, this research proposes a context-aware deep reinforcement learning (DRL) framework based on Soft Actor-Critic to enhance energy flexibility in large multizone office building during heatwaves. The approach includes historical HVAC actions and zone temperatures to capture delayed thermal responses with instantaneous states. It also leverages exogenous weather and electricity price forecasts and dynamic occupancy context. A tri-objective optimization minimizes cooling energy cost, occupant discomfort and peak power demand. Photovoltaic and thermal energy storage are accommodated to support load shifting and grid-responsive operation. The framework is evaluated using EnergyPlus-Python co-simulation with real weather and electricity market data during the 2023 Cerberus heatwave in Italy. Compared to rule-based control, it achieves a 76.55% mean reduction in occupant discomfort, 17.15% reduction in energy costs, 16.45% reduction in total energy consumption and 15.28% reduction in peak power demand. Energy flexibility metrics indicate load shifting index of 20.55% and self-consumption ratio of 95.46%. These results demonstrate the potential of context-aware DRL as a robust and scalable strategy for grid-responsive buildings under increasingly frequent and severe heatwaves.
Suggested Citation
Ahmed, Bilal & Ahmed, Ateeque & Kathirgamanathan, Anjukan & Guo, Jun-Liang & Zhang, Hong-Na & Li, Xiao-Bin & Qu, Kai-Yang & Li, Feng-Chen, 2026.
"Multi-objective context-aware deep reinforcement learning for energy flexible control of a large multizone office building under extreme heatwave conditions,"
Energy, Elsevier, vol. 360(C).
Handle:
RePEc:eee:energy:v:360:y:2026:i:c:s0360544226020347
DOI: 10.1016/j.energy.2026.141927
Download full text from publisher
As the access to this document is restricted, you may want to
for a different version of it.
Corrections
All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:eee:energy:v:360:y:2026:i:c:s0360544226020347. See general information about how to correct material in RePEc.
If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.
We have no bibliographic references for this item. You can help adding them by using this form .
If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: Catherine Liu (email available below). General contact details of provider: http://www.journals.elsevier.com/energy .
Please note that corrections may take a couple of weeks to filter through
the various RePEc services.