Bounded strategic reasoning explains crisis emergence in multi-agent market games

A version of this work was presented at the Summer Institute on Bounded Rationality 2022, at the Max Planck Institute in Berlin, Germany.

Multi-agent Market Games

This work follows along two key works. The Quantal Hierarchy model of reasoning, and the BRATS approach to market modelling.

How to run

Setup the market model

from src.market import Market
market = Market(N=100, c=0.6) # Initialise the basic model

Then to run for T timesteps use

market.run(T=1000)

Alternatively, run for an individual tick with

market.tick()

Finally, to obtain the results dictionary call

results = market.results()

For an example of how to run this work for multiple runs, see run.py

Citation

To cite the market model, please use

@article{evans2022bounded,
  title={Bounded strategic reasoning explains crisis emergence in multi-agent market games},
  author={Evans, Benjamin Patrick and Prokopenko, Mikhail},
  journal={arXiv preprint arXiv:2206.05568},
  year={2022},
  doi = {10.48550/ARXIV.2206.05568},
}

To cite the reasoning method (Quantal Hierarchy), please use

@article{evans2021bounded,
  title={Bounded rationality for relaxing best response and mutual consistency: The Quantal Hierarchy model of decision-making},
  author={Evans, Benjamin Patrick and Prokopenko, Mikhail},
  journal={arXiv preprint arXiv:2106.15844},
  year={2021},
  doi = {10.48550/ARXIV.2106.15844},
}