Skip to content

keithbadger/Fast-Policy-Iteration-for-Markov-Games-and-Robust-MDPs

Repository files navigation

Fast-Policy-Iteration-for-Markov-Games-and-Robust-MDPs

This repository stores the code used to generate the emprical results in https://arxiv.org/abs/2508.06661. The implementations of all the algorithms benchmarked are located in game_algorithm_definition.jl and rmdp_algorithm_definition.jl.

Plotting Counter-Example

To plot the counter-example, run plot_counter.jl. If you need to regenerate the function outputs, uncomment the section at the top.

Plotting Algorithms Worst Runs

The slowest runs for each algorithm can be plotted by running plot_results.jl

Generating Table of Median Runtimes

The table of median runtimes for each algorithm can be generated by running generate_tables.jl

Computing A New Data-set

To generate a new data-set for run benchmark_games.jl and benchmark_rmdps.jl for Markov games and robust MDPs respectively. Note benchmarks are multi-threaded and can be sped up by running them with more threads.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages