marl

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

🦁 A research-friendly codebase for fast experimentation of multi-agent reinforcement learning in JAX

Multi-Agent Reinforcement Learning with JAX

Fine-tuned MARL algorithms on SMAC (100% win rates on most scenarios)

BenchMARL is a library for benchmarking Multi-Agent Reinforcement Learning (MARL). BenchMARL allows to quickly compare different MARL algorithms, tasks, and models while being systematically grounded in its two core tenets: reproducibility and standardization.

VMAS is a vectorized differentiable simulator designed for efficient Multi-Agent Reinforcement Learning benchmarking. It is comprised of a vectorized 2D physics engine written in PyTorch and a set of challenging multi-robot scenarios. Additional scenarios can be implemented through a simple and modular interface.

Projects from basic algorithms to MARL. Implements MADDPG,MATD3,MA/HAPPO in Predator-Prey pursuit games with PettingZoo MPE environments.

Multi-Agent Reinforcement Learning (MARL) papers with code

Multi-Agent Reinforcement Learning (MARL) papers

[ICLR-2025] POGEMA stands for Partially-Observable Grid Environment for Multiple Agents. This is a grid-based environment that was specifically designed to be flexible, tunable and scalable. It can be tailored to a variety of PO-MAPF / MAPF settings.

A Collection of Multi-Agent Reinforcement Learning (MARL) Resources

A framework for creating rich, 3D, Minecraft-like single and multi-agent environments for AI research. (Accepted at ICML 2025).

Open-Source Library for Fully Cooperative Multi-LLM Reinforcement Learning

A custom MARL (multi-agent reinforcement learning) environment where multiple agents trade against one another (self-play) in a zero-sum continuous double auction. Ray [RLlib] is used for training.

This is a framework for the research on multi-agent reinforcement learning and the implementation of the experiments in the paper titled by ''Shapley Q-value: A Local Reward Approach to Solve Global Reward Games''.

Training of Drone Swarms using StableBaselines3, PettingZoo, AirSim and UE4

无人机动态覆盖控制；1. 实现了一个无人机点覆盖环境；2. 给出了无人机连通保持规则；3. 给出了基于MARL的控制算法

[NeurIPS 2021] CDS achieves remarkable success in challenging benchmarks SMAC and GRF by balancing sharing and diversity.

A tool for aggregating and plotting MARL experiment data.

A solution for Dynamic Spectrum Management in Mission-Critical UAV Networks using Team Q learning as a Multi-Agent Reinforcement Learning Approach