Empirical Game-Theoretic Analysis for Credit Networks

This project applies empirical game-theoretic analysis (EGTA) for credit networks. Currently, the code adheres to the prepayment operation.

Code Structure

classic_EGTA

1. nash_solvers: This folder contains replicator dynamics and gambit for analyzing games.
2. clearing.py: Functions for clearing a credit network.
3. egta_solver.py: The main solver class for EGTA, including simulating profiles and equilibrium computation.
4. evaluation.py: Functions for analyzing the equilibria and computing price of anarchy.
5. net_generator.py: Functions for generating instances of credit networks.
6. player_reduction.py: Functions for player reduction.
7. strategies: Definitions for heuristic strategies.
8. symmetric_utils.py: Functions for creating compact representation for symmetric games, converting a compact representation to complete payoff matrix, and computing pure-strategy equilibria.

envs

1. prepayment_net: A Pettingzoo environment for credit networks with prepayment.

instances

This folder includes instances generated with different network parameters.

example.py

An example for running EGTA for credit networks.