Our program has been tested to work on Python 3.12-3.13, and may not work on older versions.
First, install necessary requirements with pip install -r requirements.txt.
The model can be trained, or the best result from a saved model can be run.
The config file contains parameters that can be easily modified. Important parameters to note for running the best result are:
RUN_WITH_GRAPHICSGA_MODEL_FILE(if it was changed)
The rest of the parameters are mainly used for learning and should not be changed.
run_best_result.py is used to, as the name suggests, run the best result.
This loads the best result from the GA_MODEL_FILE using pygad, and then runs
our controller with the best chromosome in that saved state. This means that you
must have pygad installed (it is in requirements.txt after all) to run our controller.
If our controller must be run in a different way, other files in the src directory
must still be included anyway as the controller may use them for type hints or various
other things. In theory our controller can be run without pygad by manually entering
the list of floats best chromosome that is stored in the results in the GA_MODEL_FILE.
To train the model, run genetic_learner.py, making sure to first set flags appropriately
in the config file , such as disabling graphics and setting a reasonable number of
processes for your hardware.
If you have any issues, please email me at [email protected].
- Cedric Boucher
- Aiden Teal
- Mohamed Alzinnad