Algorithmic collusion in double-sided markets: A rideshare example

The paper related to this repo has been submitted to ITSC 2024 and is under review. The PPO implementation used has been taken from https://github.com/seolhokim/Mujoco-Pytorch

Dependencies

• pytorch
• pettingzoo
• wandb

Use pip install -r requirements.txt

Instructions

Configuration parameters for the PPO algorithm can be found in the config.ini OD matrix and the cost matrix can be directly changed in the run-custom.py Executing run-custom.py will start the simulation and progress can be tracked in wandb. Use the following to execute simulation: python run-custom.py