Realizability guaranteed multi-timescale optimization decision for home energy management

A Home Energy Management (HEM) optimization decision problem can be modeled with multiple timescales. In this paper, a two-timescale deterministic optimization decision for Heating, Ventilation and Air Conditioning (HVAC) of HEM is investigated, in order to minimize the deviation from indoor temperature and user desired temperature with constraint of energy consumption. The two timescales are slow timescale and fast time scale which are related, so there is a challenge that the consistency of these timescales should be guaranteed. An induction method for building the two-timescale model is proposed, that is, the fast timescale model is built at first, and then the slow timescale model is induced from the fast timescale model. Therefore the method can guarantee the realizability of fast timescale decision under slow timescale decision. And we prove that the existing empirical models cannot guarantee the consistency of models in two timescales, so it leads to the problem of realizability. Simulation results show that the proposed model guarantees the consistency of the fast timescale model and the slow timescale model, and the model can satisfy user expectation with better performance.