PPO-PackCooling

Forked from PPO-PyTorch by Nihkil Baharte

Source code for the paper "Actor-Critic Methods using Physics-Informed Neural Networks: Control of a 1D PDE Model for Fluid-Cooled Battery Packs" by Amartya Mukherjee and Jun Liu, 2023

Implementation of HJBPPO on HJBPPO.py

Implementation of the PackCooling environment is in environments/PackCooling.py

Implementation of the rendering feature in the PackCooling environment is in environments/PackCoolingGraph.py

To train HJBPPO on the environment, set the prob_optimal_control parameter in HJBPPO.HJBPPO to 0.5 To train HJB value iteration on the environment, set the prob_optimal_control paper in HJBPPO.HJBPPO to 1.0

References

• PPO paper
• OpenAI Spinning up
• PPO pack cooling paper