03

Reinforcement Learning Thermostat Controller Comparison

Simulated thermal-control study comparing Q-learning, SARSA, DQN, rule-based, and proportional thermostat controllers under matched conditions for temperature error, normalized control effort, reward, and action switching.

Role
Validation & Documentation
Stack
Python · NumPy · Matplotlib · PyTorch · Reinforcement Learning
View full project

Final controller comparison

Temperature trajectories for Q-learning, SARSA, DQN, rule-based, and proportional controllers under shared simulated conditions with a 25 degree Celsius target.
Average simulated temperature error comparison for the five main thermostat controllers.
Average normalized control effort for the five main thermostat controllers.
Average action-switching comparison showing more frequent switching by the learned thermostat controllers than the traditional baselines.
010203040506

Engineering

Project archive

03

Reinforcement Learning Thermostat Controller Comparison

Simulated thermal-control study comparing Q-learning, SARSA, DQN, rule-based, and proportional thermostat controllers under matched conditions for temperature error, normalized control effort, reward, and action switching.

Role
Validation & Documentation
Stack
Python · NumPy · Matplotlib · PyTorch · Reinforcement Learning
View full project

Final controller comparison

Temperature trajectories for Q-learning, SARSA, DQN, rule-based, and proportional controllers under shared simulated conditions with a 25 degree Celsius target.
Average simulated temperature error comparison for the five main thermostat controllers.
Average normalized control effort for the five main thermostat controllers.
Average action-switching comparison showing more frequent switching by the learned thermostat controllers than the traditional baselines.