This repo contains the Python code used to run simulations for the Mean Field Equilibria for Multiarmed Bandit Games paper. A preprint can be found on SSRN.
- Python 3
- Jupyter notebooks
- MCSim: small framework for Monte Carlo simulations
- Toyplot: library used to create plots
- Statsmodels: statistics library used to fit LOWESS
High level code to run simulations and plot results are Jupyter notebooks:
simulation_playground
: this notebook gives a simple example of how to set up a particular simulation, and useful if you want to simulate a specific scenariorun_sims_disk
: this notebooks runs the simulations found in the paper and saves the outcomes as pickle objects. Before using this notebook, runsetup.sh
to create subdirectories used to store simulations and plots.plots
: this notebook recreates the plots used in the paper. It reads the pickle files created by therun_sims_disk
notebook, so run that first.
If you have any questions about the code or simulations, or have trouble getting things to work, please open an issue.