Using Muzero algorithm to design toolpaths for additive manufacturing process given any section geometry. Currently, we use a dense reward structure where a positive reward is assigned to correct material deposition, a negative reward is assigned to wrong material deposition, and a small negative reward is assigned to other motions at each time step. Significant parts of this code are adopted based on muzero-general.
- Accelerate the code
- Dense reward structure test
- Sparse reward structure test
- Document hyper-parameter analysis
- Pretraining networks
git clone https://github.com/mojtabamozaffar/toolpath-design-rl
cd toolpath-design-rl
pip install -r requirements.txt
python main.py
This project is released under the MIT License.