# openai/maddpg

Code for the MADDPG algorithm from the paper "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments"

Repository: https://github.com/openai/maddpg
Canonical: https://ross.abutalabs.com/products/maddpg
Homepage: https://arxiv.org/pdf/1706.02275.pdf
Language: Python
License: MIT
License Family: permissive
Topics: paper
Archived: true
Last push: 2024-04-01T21:04:21+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3135, "days_push": 884, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1981, forks 534 (observed 2026-08-28T04:06:01.730103+00:00)

## What it is
Reference implementation of the MADDPG (Multi-Agent Deep Deterministic Policy Gradient) algorithm from OpenAI's paper on mixed cooperative-competitive multi-agent environments. It is a Python library designed to train agents in the Multi-Agent Particle Environments (MPE) using TensorFlow.

## Use cases
- reproduce the MADDPG paper results
- train multi-agent reinforcement learning policies
- compare MADDPG against DDPG in cooperative-competitive settings
- experiment with multi-agent particle environments
- study centralized training with decentralized execution

## When to choose
- you need the original reference implementation of MADDPG for research or study
- you want to train agents in OpenAI's Multi-Agent Particle Environments
- you are comparing multi-agent actor-critic algorithms

## When to avoid
- you need a maintained library with modern TensorFlow or PyTorch support
- you want production-ready multi-agent RL tooling
- you need up-to-date dependencies or bug fixes

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, reinforcement-learning
- domain: machine-learning, artificial-intelligence
- platform: python
- tags: maddpg, multi-agent, actor-critic, deep-learning, paper-code, openai, mpe, research

## Member repositories
- openai/maddpg (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:01.730103+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:03:59.070188+00:00, confidence not recorded.
  - readme: https://github.com/openai/maddpg (fetched 2026-08-28T04:06:01.730103+00:00, sha a267b6be22a9)
- Data as of 2026-08-30T08:39:29.467469+00:00.