In this repository we provide all necessary code to reproduce the experiments included in our paper "Gradient Methods for solving Stackelberg Games", presented to the 2019 Algorithmic Decision Theory conference.
The jupyter notebook notebooks/easy_example.ipynb reproduces the experimento of Section 4.1.
In order to reproduce the experiment of Figure 3 (a) of Section 4.2, run python src/paper_big_experiment.py. This script will write the results in ´results/exp1´. To analyse these results, the notebook notebooks/analysis_adv_reg.ipynb was used.
In order to reproduce the experiment of Figure 3 (b) of Section 4.2, run python src/convergence_experiment.py. Results will be written in results/exp_convergence. The notebook notebooks/analysis_adv_rer_convergence.ipynb was used to analyse such results.