Optimal Control of Multiroom HVAC System: An Event-Based Approach.

Building energy saving is of great practical interest due to the increasing energy consumption in buildings. The optimal control of the heating, ventilation, and air-conditioning (HVAC) systems leads to great energy saving potential. However, this problem is challenging due to the exponentially increasing state space and policy space. In this brief, we consider this important problem and make the following major contributions. First, we formulate the multiroom HVAC control problem as an event-based optimization, where decisions are made only when certain events occur. The size of the event space is significantly smaller than that of the state space. Second, to further simplify the calculation process, we develop an approximate solution method which focuses on local-event-based policies. These policies control the terminal devices in a room using solely the information in that room. Third, we demonstrate the performance of this method through two sets of numerical examples. In the small-scale two-room example, it is shown that our method can achieve a near-optimal solution. In the large-scale example, it is shown that the local-event-based approach can achieve a policy which is better than the threshold-based control method, hysteresis control method, and predictive control method.