MARL-DPP is an open-source framework enabling multi-agent reinforcement learning (MARL) with enforced diversity through Determinantal Point Processes (DPP). Traditional MARL approaches often suffer from policy convergence to similar behaviors; MARL-DPP addresses this by incorporating DPP-based measures to encourage agents to maintain diverse action distributions. The toolkit provides modular code for embedding DPP in training objectives, sampling policies, and managing exploration. It includes ready-to-use integration with standard OpenAI Gym environments and the Multi-Agent Particle Environment (MPE), along with utilities for hyperparameter management, logging, and visualization of diversity metrics. Researchers can evaluate the impact of diversity constraints on cooperative tasks, resource allocation, and competitive games. The extensible design supports custom environments and advanced algorithms, facilitating exploration of novel MARL-DPP variants.

Multi-Agent Miners offers a grid-world environment where multiple autonomous miner agents navigate, dig, and collect resources while interacting with each other. It supports configurable map sizes, agent counts, and reward structures, allowing users to create competitive or cooperative scenarios. The framework integrates with popular RL libraries via PettingZoo, providing standardized APIs for reset, step, and render functions. Visualization modes and logging support help analyze behaviors and outcomes, making it ideal for research, education, and algorithm benchmarking in multi-agent reinforcement learning.

This repository provides a complete suite of multi-agent reinforcement learning algorithms—including MADDPG, DDPG, PPO, and more—integrated with standard benchmarks like the Multi-Agent Particle Environment and OpenAI Gym. It features customizable environment wrappers, configurable training scripts, real-time logging, and performance evaluation metrics. Users can easily extend algorithms, adapt to custom tasks, and compare policies across cooperative and adversarial settings with minimal setup.

Scalable MADDPG is a research-oriented framework for multi-agent reinforcement learning, offering a scalable implementation of the MADDPG algorithm. It features centralized critics during training and independent actors at runtime for stability and efficiency. The library includes Python scripts to define custom environments, configure network architectures, and adjust hyperparameters. Users can train multiple agents in parallel, monitor metrics, and visualize learning curves. It integrates with OpenAI Gym-like environments and supports GPU acceleration via TensorFlow. By providing modular components, Scalable MADDPG enables flexible experimentation on cooperative, competitive, or mixed multi-agent tasks, facilitating rapid prototyping and benchmarking.

AutoDRIVE Cooperative MARL is a GitHub-hosted framework combining the AutoDRIVE urban driving simulator with adaptable multi-agent reinforcement learning algorithms. It includes training scripts, environment wrappers, evaluation metrics, and visualization tools to develop and benchmark cooperative driving policies. Users can configure agent observation spaces, reward functions, and training hyperparameters. The repository supports modular extensions, enabling custom task definitions, curriculum learning, and performance tracking for autonomous vehicle coordination research.

DeepMind MAS Environment is a Python library that provides a standardized interface for building and simulating multi-agent reinforcement learning tasks. It allows users to configure number of agents, define observation and action spaces, and customize reward structures. The framework supports agent-to-agent communication channels, performance logging, and rendering capabilities. Researchers can seamlessly integrate DeepMind MAS Environment with popular RL libraries such as TensorFlow and PyTorch to benchmark new algorithms, test communication protocols, and analyze both discrete and continuous control domains.

MADDPG-Keras delivers a complete framework for multi-agent reinforcement learning research by implementing the MADDPG algorithm in Keras. It supports continuous action spaces, multiple agents, and standard OpenAI Gym environments. Researchers and developers can configure neural network architectures, training hyperparameters, and reward functions, then launch experiments with built-in logging and model checkpointing to accelerate multi-agent policy learning and benchmarking.

Selective Reincarnation introduces a dynamic population-based training mechanism tailored for multi-agent reinforcement learning. Each agent’s performance is regularly evaluated against predefined thresholds. When an agent’s performance falls below its peers, its weights are reset to those of the current top performer, effectively reincarnating it with proven behaviors. This approach maintains diversity by only resetting underperformers, minimizing destructive resets while guiding exploration toward high-reward policies. By enabling targeted heredity of neural network parameters, the pipeline reduces variance and accelerates convergence across cooperative or competitive multi-agent environments. Compatible with any policy gradient-based MARL algorithm, the implementation integrates seamlessly into PyTorch-based workflows and includes configurable hyperparameters for evaluation frequency, selection criteria, and reset strategy tuning.

CrewAI-Learning is an open-source library designed to streamline multi-agent reinforcement learning projects. It offers environment scaffolding, modular agent definitions, customizable reward functions, and a suite of built-in algorithms such as DQN, PPO, and A3C adapted for collaborative tasks. Users can define scenarios, manage training loops, log metrics, and visualize results. The framework supports dynamic configuration of agent teams and reward sharing strategies, making it easy to prototype, evaluate, and optimize cooperative AI solutions across various domains.