A scenario-based distributed stochastic MPC for building temperature regulation

Yushen Long, Shuai Liu, Lihua Xie, Karl Henrik Johansson

Abstract: In this paper, we focus on the temperature regulation of rooms in buildings. By using the dynamic model of the thermal process, weather condition, occupancy and so on, a Stochastic Model Predictive Control (SMPC) problem is formulated to keep the temperature of rooms within a comfortable range with a predefined probability while consuming less energy. The temperature regulation problem in this paper is an optimal control problem of a linear system with additive uncertainty. To overcome the computational burden caused by the large number of rooms, a subgradient-based dual decomposition method is used to solve the SMPC problem in a distributed manner. Simulation results show the effectiveness of our results.