Awesome GitHub RepositoriesGame Simulation Environments

This project is a reinforcement learning toolkit and simulation-based AI trainer for creating intelligent agents within Unity simulations. It provides a multi-agent simulation framework for configuring cooperative or competitive scenarios and includes an environment wrapper that bridges simulations with standard machine learning libraries using gym-style interfaces. The system features a native cross-platform inference engine that executes trained neural network models for real-time decision making without external dependencies. It enables the acceleration of the learning process by running m

This project is a comprehensive framework for building and managing autonomous agent systems. It provides a unified architecture for orchestrating multi-agent societies, where specialized agents collaborate through roleplay to decompose and solve complex tasks. The system integrates language models with external environments, enabling agents to perform real-world actions through a standardized tool-calling abstraction layer. The framework distinguishes itself through its focus on iterative reasoning and data reliability. It employs automated feedback loops to refine agent outputs and self-eva

Grasscutter is a private game server emulator designed to replicate game backend logic and simulate core gameplay mechanics without the use of official servers. It functions as a system for simulating character progression, world entities, and general backend simulation to host a private instance of a specific anime game. The project includes a game object management console for spawning entities and controlling player inventories within the simulated environment. It also provides a server configuration tool featuring a guided interface to generate and manage the settings and configuration fi

ArchiSteamFarm is a background service designed for the automated management of multiple Steam accounts. It functions as a centralized platform that coordinates account settings, inventory organization, and game idling tasks across several profiles from a single interface. The software utilizes a state-machine approach to track individual account lifecycles and communicates with the Steam platform through an event-driven protocol. It supports configuration-driven logic injection, allowing users to adjust operational parameters without modifying the source code, and provides a plugin-based int

pysc2 es una interfaz de Python y un framework de simulación que conecta el motor del juego StarCraft II con agentes de aprendizaje automático. Actúa como un envoltorio de API que expone los componentes internos del juego como un conjunto de observaciones y acciones, proporcionando un entorno de aprendizaje por refuerzo para la investigación y el entrenamiento. El framework incluye herramientas para el análisis de repeticiones de juegos para extraer datos y secuencias de acciones de partidas grabadas para el modelado predictivo. También proporciona un entorno de simulación de agentes para ejecutar y evaluar el rendimiento de agentes de inteligencia artificial individuales o competitivos. El sistema maneja la configuración del mapa del juego, el análisis de comportamiento y proporciona una interfaz de control manual para depurar el comportamiento del agente. Transforma el estado del juego en tensores multidimensionales y utiliza un framework de llamada a procedimiento remoto para gestionar la comunicación entre el cliente y el motor del juego.

A/B Street is an open-source traffic simulation and urban planning tool that models how cars, bikes, and pedestrians move through real-world street networks. It imports data from OpenStreetMap to build detailed, lane-level road models, then runs discrete-event simulations to analyze travel times, delays, and congestion patterns across different infrastructure scenarios. The project provides an interactive map editor for modifying road geometry, lane configurations, traffic signals, and access restrictions, with full undo/redo support. Users can design low-traffic neighborhoods by placing moda

OpenTTD is an open-source game engine and transport simulation game. It provides an isometric sandbox environment for building and managing complex logistics and transport networks. The project functions as a multiplayer simulation sandbox where users can build infrastructure cooperatively or competitively in a shared virtual world. The platform is designed as a moddable simulation system that supports external assets, graphics, and gameplay modifications. It includes mechanisms for downloading and integrating add-on content and utilizes a plugin-based system to extend game mechanics beyond t

Open Spiel es una biblioteca de investigación y framework para aprendizaje por refuerzo, planificación y simulación de juegos multi-agente. Proporciona un sistema para representar juegos de agente único y multi-agente en escenarios de suma cero, cooperativos y de información imperfecta. El proyecto utiliza una interfaz de juego abstracta estandarizada para desacoplar la lógica del juego de los algoritmos, permitiendo que los agentes se ejecuten en diferentes tipos de juegos. Implementa lógica crítica para el rendimiento en C++ con bindings para Python y utiliza simulación determinista con semillas para garantizar la reproducibilidad en la investigación de benchmarks. El framework cubre una amplia gama de capacidades, incluyendo algoritmos de búsqueda y planificación computacional, aprendizaje por refuerzo multi-agente e investigación en teoría de juegos. También incluye herramientas para el análisis de dinámicas de aprendizaje para rastrear el comportamiento de los agentes y calcular métricas de evaluación.

DouZero es un framework de aprendizaje por refuerzo profundo y sistema de entrenamiento diseñado para enseñar a agentes digitales a dominar juegos de cartas complejos. Proporciona la infraestructura para implementar pipelines de aprendizaje por refuerzo de alto rendimiento y evaluar el éxito competitivo de los agentes de juego. El sistema utiliza una arquitectura de actor-aprendiz distribuida que separa los actores de simulación de juego de los dispositivos de entrenamiento de GPU para acelerar la convergencia del modelo. Combina la búsqueda de árbol de Monte Carlo con la estimación de valor basada en políticas para determinar los movimientos óptimos a través de evaluación recursiva y muestreo aleatorio. El kit de herramientas incluye capacidades para bucles de simulación de auto-juego para generar datos de entrenamiento sintéticos y buffers de experiencia asíncronos para entrenamiento por lotes muestreados. También cuenta con herramientas de benchmarking para medir el rendimiento del agente comparando tasas de victoria contra datasets aleatorios, basados en reglas o humanos.

Malmo is a voxel-based simulation platform designed for artificial intelligence research and the study of autonomous agent behaviors. Built as a sandbox environment using Minecraft, it serves as a framework for multi-agent simulation and reinforcement learning research within a 3D grid of blocks. The project distinguishes itself through a multi-agent simulation framework that coordinates and synchronizes multiple autonomous agents to perform collaborative missions. It provides a standardized interface following reinforcement learning specifications, allowing it to function as an environment f

SUMO is a microscopic traffic simulator and road network generator designed to model the detailed movement of individual vehicles, pedestrians, and public transport. It serves as a multimodal transport simulator and an autonomous vehicle simulator, enabling the analysis of interactions between different transport modes and the testing of automated driving behaviors. The project distinguishes itself through its ability to couple simulations with external controllers and network simulators to model vehicle communication and autonomous control transitions. It supports the creation of simulation-

Minigo is a TensorFlow-based reinforcement learning engine designed to master the game of Go. It functions as a comprehensive system for training neural networks to predict board policies and game outcomes, utilizing a model trainer to generate self-play data and optimize weights. The project is distinguished by its ability to perform large-scale game simulations using Kubernetes to distribute worker nodes across CPU, GPU, and TPU hardware. It employs a Monte Carlo Tree Search implementation to identify optimal moves and supports specialized hardware acceleration, including inference on Edge

Lingbot-world is an interactive world simulator and framework for generating high-fidelity video environments from text and image prompts. It functions as a video generation system designed to create controllable simulations for applications such as robotics learning and gaming. The project includes a video motion controller that directs camera and object movement using transformation matrices and action strings. It utilizes a quantized inference engine to reduce memory usage and accelerate the generation of video sequences. The system covers a range of optimization techniques, including fou