Smart-Model-then-Control

This work innovates traditional model predictive control for the scheduling of thermostatically controlled loads. Inspired by "Smart Predict-then-Optimize", a "Smart Model-then-Control" strategy is proposed to learn a cost-effective model for the downstream control task. The actual control costs are reduced in multiple building types.

Codes for submitted Paper "A Smart Model-then-Control Strategy for the Scheduling of Thermostatically Controlled Loads".

Authors: Xueyuan Cui, Boyuan Liu, Yehui Li, and Yi Wang.

Requirements

Python version: 3.8.17

The must-have packages can be installed by running

pip install requirements.txt

Experiments

Data

All the data for experiments can be downloaded from Google Drive.

Reproduction

To reproduce the experiments of the proposed methods and comparisons for single-zone, 22-zone, and 90-zone buildings, please go to folders

cd #Codes/Single-zone
cd #Codes/22-zone
cd #Codes/90-zone

respectively. The introduction on the running order and each file's function is explained in Readme.md in the folder.

Note: There is NO multi-GPU/parallelling training in our codes.

The required models as the warm start of SMC are saved in #Results.

Citation