CITATION.cff

cff-version: 1.2.0
message: "If you use EconGym in your research, please cite it as below."
title: "EconGym: a Gymnasium-style suite of economics reinforcement-learning environments with closed-form equilibrium benchmarks"
type: software
authors:
  - family-names: Lamba
    given-names: Shubh
version: 0.2.0
license: MIT
repository-code: "https://github.com/shubhL-research/econgym"
url: "https://github.com/shubhL-research/econgym"
keywords:
  - reinforcement-learning
  - algorithmic-collusion
  - bertrand-competition
  - industrial-organization
  - gymnasium
  - pettingzoo
  - economics
abstract: >-
  EconGym is a dependency-light (numpy-only by default) suite of Gymnasium- and
  PettingZoo-style economics environments, each paired with a closed-form
  equilibrium benchmark as the single source of truth. It ships a byte-for-byte
  faithful reproduction of the Q-learning Bertrand-collusion dynamics of "The
  Two Faces of Algorithmic Collusion", alongside Cournot, differentiated
  Bertrand, first/second-price auctions, public goods, repeated Prisoner's
  Dilemma, and Rubinstein bargaining. An optional interop layer exposes the
  environments through real Gymnasium/PettingZoo APIs for use with standard RL
  tooling.
identifiers:
  - type: doi
    value: 10.5281/zenodo.20788510
    description: Zenodo archive