Choosing the Best Strategy for Energy Aware Building System: an SVM-based Approach

For many old buildings in the world, due to the legacy devices problem, it is hard to supply appropriate energy for them. In order to reduce the energy consumption of buildings under the premise of satisfying user requirements, we use software control systems whose core part is the scheduling strategy, to reconstruct them. It is time consuming to choose a good scheduling strategy due to many uncertain factors, among which user actions are of the most influence. In this paper, we propose an Support Vector Machine (SVM) based approach to explore the relation between user action and the best scheduling strategy of a control system. The main contributions include: (1) obtaining the sample set by collecting data at the model level using Statistical Model Checking (SMC) based method; (2) using SVM algorithm to learn the relation model between user actions and the best scheduling strategies; and (3) applying the relation model to predict a best scheduling strategy. Finally a real case study is conducted showing the efficiency of our approach. Keywords—Energy Aware Building; User Actions; Scheduling Strategy; Support Vector Machine; Statistical Model Checking.