This repository implements an AI-powered Energy Consumption Optimization System using Deep Reinforcement Learning (DQN).
The goal is to reduce cooling energy usage while maintaining stable indoor temperature.
Key components included in this project:
- A DQN agent trained to minimize energy usage
- A custom RL environment simulating HVAC-style temperature dynamics
- Training and evaluation utilities
- A Streamlit dashboard for real-time simulation
- Deep Q-Learning agent with neural network function approximation
- Custom temperature + energy reward system
- Modular RL pipeline (environment, agent, training, evaluation)
- Streamlit UI for demonstration and visualization
- Easily extensible project structure
Environment (HVAC)
- Temperature simulation
- Energy penalty calculation
Outputs → state, reward
DQN Agent (Neural Network)
- Predicts Q-values
- Selects best action
Energy Control System
- Applies cooling power based on chosen action
Clone the repository: git clone https://github.com/aanishraj777/Energy-Optimization-Using-AI.git cd Energy-Optimization-Using-AI
Install dependencies: pip install -r requirements.txt
Run training: python src/train.py
Outputs:
- Trained model saved in
models/ - Training plots saved in
results/training_plots/
Run evaluation: python src/evaluate.py
Outputs:
- Temperature comparison graph
- Energy usage comparison graph
- Metrics saved in results directory
Start the dashboard: streamlit run app.py
This launches an interactive simulation comparing:
- AI-controlled temperature
- Baseline temperature
- Energy usage differences
Run all tests: pytest tests/
- Python
- TensorFlow / Keras
- NumPy
- Matplotlib
- Streamlit
- PyYAML
- Multi-agent reinforcement learning
- Deploy Streamlit UI online
- Integrate live IoT sensor data
- Hyperparameter optimization (Optuna / Ray Tune)
Pull requests are welcome!
Fork → Branch → Commit → Pull Request.
If you find this useful, please consider giving the repository a star ⭐ on GitHub!