RL-based-event-triggered-MPC

This is official implementation of our paper: Event-Triggered Model Predictive Control with Deep Reinforcement Learning.

Installation

• create a anaconda environment via: conda create -n empc python=3.8 -y
• activate the virtual env via: conda activate empc
• install the requirements via: pip install -r requirements.txt

Algorithms

All the rule-based/RL algorithms included in our RL-eMPC framework.

• Threshold-based.
• LSTDQ.
• DQN.
• DDQN.
• DQN/DDQN + PER (prioritized experience buffer) + LSTM
• A2C (Discrete)
• PPO (Discrete)
• SAC (Discrete)

To visualize the training process, you can run: tensorboard --logdir runs, for instance:

References

The A2C, PPO, and SAC code are based on the following wonderful repos, please give the credits to the authors.

• A2C
• PPO
• SAC

Cite

Please consider cite our paper if you find this repo is useful. https://urldefense.com/v3/__http://arxiv.org/abs/2208.10302__;!!HXCxUKc!2bA5zV6LbesRBR_dmzZQ_96creDx9NU6EPGTNBMnAGMxbySbWfm86Qz07MCOAnk5yK9hmxiO7Of5UxU9Pg$