Cumincad : CUMINCAD Papers : Paper caadria2005_b_4a_a:Reinforcement Learning and Real-time Building Thermal Performance Data Visualization

id	caadria2005_b_4a_a
authors	Ravi S. Srinivasan, Ali M. Malkawi
year	2005
title	Reinforcement Learning and Real-time Building Thermal Performance Data Visualization
source	CAADRIA 2005 [Proceedings of the 10th International Conference on Computer Aided Architectural Design Research in Asia / ISBN 89-7141-648-3] New Delhi (India) 28-30 April 2005, vol. 2, pp. 141-148
doi	https://doi.org/10.52842/conf.caadria.2005.141
summary	Computational Fluid Dynamics (CFD) simulations are used to predict the fluid behavior and particle systems-in-action in three-dimensional space, allowing experts to evaluate a series of environmental decisions in designing buildings. Although computing power has increased in the past decade, detailed CFD simulations introduce time-delay that defeats the notion of real-time data visualization. A method that can bypass the time-consuming simulations and generate results comparable to detailed CFD simulations will allow such visualizations to be constructed. This paper discusses a pilot project that utilizes a Reinforcement Learning (RL) algorithm coupled with a simplified fluid dynamics equation to generate thermal performance data for real-time visualization.
series	CAADRIA
email
full text	file.pdf (64,555 bytes)
last changed	2022/06/07 08:00