RL for Goal Recognition

Supplementary material for AAAI2022. This packages include the code to generate the results of our paper, besides the R&G approach. The R&G approach is not included in this package, only the necessary code to run the developed methods of our paper.

Setting up the environment

This package was tested on Python 3.6.13 and should work on Python 3.6 or higher. To install requirements (works on virtualenv and conda), run:

pip install -r requirements.txt

Running experiments

To compute the results for partial observability (0.1,0.3,0.5,0.7,1.0), run the following command inside the src/ folder:

python experiments.py

The results are going to be outputed on src/results.txt. To run the experiments with noise, run the following command inside the src/ folder:

python experiments_noisy.py

The results are going to be outputed on src/results_noisy.txt.