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.

The MARL Simulator is designed to facilitate efficient and scalable development of multi-agent reinforcement learning (MARL) algorithms. Leveraging PyTorch's distributed backend, it allows users to run parallel training across multiple GPUs or nodes, significantly reducing experiment runtime. The simulator offers a modular environment interface that supports standard benchmark scenarios—such as cooperative navigation, predator-prey, and grid world—as well as user-defined custom environments. Agents can utilize various communication protocols to coordinate actions, share observations, and synchronize rewards. Configurable reward and observation spaces enable fine-grained control over training dynamics, while built-in logging and visualization tools provide real-time insights into performance metrics.

MARTI (Multi-Agent Reinforcement learning Toolkit and Interface) is a research-oriented framework that streamlines the development, evaluation, and benchmarking of multi-agent RL algorithms. It offers a plug-and-play architecture where users can configure custom environments, agent policies, reward structures, and communication protocols. MARTI integrates with popular deep learning libraries, supports GPU acceleration and distributed training, and generates detailed logs and visualizations for performance analysis. The toolkit’s modular design allows rapid prototyping of novel approaches and systematic comparison against standard baselines, making it ideal for academic research and pilot projects in autonomous systems, robotics, game AI, and cooperative multi-agent scenarios.

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.