Phyiscs-informed_Reinforcement_Learning

Novel Physics-informed approach to model-free RL for physics-based (PDE) control

Jupyter notebooks used for experiementation are included.

See thesis for more detailed explanation.

Currently working to submit thesis (after revision) for publication to physics & machine learning journals on supervisors reccomendation.

Name		Name	Last commit message	Last commit date
Latest commit History 15 Commits
Agents.py		Agents.py
DQN-standard_Diffusion.ipynb		DQN-standard_Diffusion.ipynb
DQN_Diffusion.ipynb		DQN_Diffusion.ipynb
DQN_Diffusion_PINN_3D.ipynb		DQN_Diffusion_PINN_3D.ipynb
DQN_Diffusion_PINN_3D_exp.ipynb		DQN_Diffusion_PINN_3D_exp.ipynb
PI-DQN_Diffusion.ipynb		PI-DQN_Diffusion.ipynb
PINN_3D.py		PINN_3D.py
PI_Reinforcement_Learning_Thesis.pdf		PI_Reinforcement_Learning_Thesis.pdf
README.md		README.md
TensorForce_PINN_convection.ipynb		TensorForce_PINN_convection.ipynb
TensorForce_PINN_diffusion.ipynb		TensorForce_PINN_diffusion.ipynb
Tensorforce_PINN_wrapper_Convection.ipynb		Tensorforce_PINN_wrapper_Convection.ipynb
Tensorforce_PINN_wrapper_Diffusion.ipynb		Tensorforce_PINN_wrapper_Diffusion.ipynb
environment.py		environment.py
models.py		models.py
wrapper.py		wrapper.py

Provide feedback