Develop a RL tool to evaluate/test the efficiency of grid networks for transport operations.
This project uses miniconda for packaging and dependency management.
After installing miniconda for your system:
-
Create a venv and install the projects dependencies
conda env create --file environment.yml
-
Based on your OS, follow the instructions on this website to install pytorch. You can also make use of CUDA or MPS, if you install the library correctly based on your OS.
If you use MacOS, follow the instructions here to properly install pytorch and make use of cuda
You can check the parameters in ./src/options.py
file
```commandline
python run.py --problem tsp --graph_size 10 --run_name 'tsp10_rollout'
```
or
```commandline
run.py --problem cvrp --graph_size 10 --run_name 'vrp10_rollout'
```
```commandline
python generate_data.py --problem all --name validation --seed 4321
```
```commandline
tensorboard --logdir 'path/to/master-thesis-2023-reinforcement-learning-in-grids/'
```
pytest