Pythermostat

A tool to evaluate thermal controllers on a given thermal model

Controllers

1. Valve controller: Heat only below threshold temperature.
2. DQN controller: A reinforcement based controller, learning when to heat through experience.
3. Opt controller: An optimized control using MILP. Maximizing comfort and minimizing energy consumption.

Model

The model used is the dark-grey box model. More on this here and here.

Install & Getting started

Using python 3.7.9, install dependencies with pip install -r requirements.txt. Then you should be able to run the main file and get results in the results folder.

If wou want to run Opt controller, you'll need to install a solver: gurobi (or cbc, cplex, ...)

Dqn controller uses Stable-baselines librabry, you may want to install external dependencies such as mpi.