Bayesian network-based air-conditioning control considering of occupants requests

Abstract

In this paper, we propose an air-conditioning control methodology based on a prediction of occupant's request. Our method enables to predict how an occupant wants to operate switches on the air-conditioner control panel in certain thermal environment using the Bayesian Networks. To show our basic concept and detailed procedure, we first provide probability tables which presents the probabilistic distribution of occupants' requests at each temperature width. Next, we show the structure of Bayesian Networks. After that, using an example case, we describe how to predict occupants requests and how to update probability tables. Finally, we confirm our proposed method discussing our experimental results.