Awesome AI Agents › Agent Monitoring Tools

TheoLvs/westworld

⭐ 60 Jupyter Notebook added to this list on 2025-04-19 repository created 2020-04-01

Westworld is a multi-agent simulation library developed in Python designed to simulate and optimize systems and environments where multiple agents interact. Inspired by Unity software and Unity ML Agents, this library adapts similar concepts into Python to facilitate the creation and management of complex multi-agent environments. It targets applications in various domains such as logistics, retail, and epidemiology by providing pre-coded spatial environments and enabling communication between agents. The library supports optimization through heuristics and reinforcement learning techniques, making it a versatile tool for researchers and developers interested in multi-agent systems. The library is currently in an experimental alpha stage, actively under development, and users are cautioned about potential instability and incomplete documentation. Westworld offers features such as easy creation of grid and non-grid environments, various object types including agents, obstacles, collectibles, and triggers, and the ability to subclass these objects for customization. It includes a spawner for random object generation, a basic rigid body system, and simple agent behaviors like pathfinding, wandering, random walk, fleeing, and vision range. Additional functionalities include automatic maze generation, layer integration to convert images into obstacles aligned to a grid, and simulation visualization with replay and export options in GIF or video formats. The roadmap for the project includes expanding classic simulations and tutorials, improving pathfinding algorithms, integrating reinforcement learning more seamlessly with Stable Baselines, and enhancing visualization capabilities for web integration beyond PyGame. Installation is straightforward via PyPi, and the project supports development through Poetry for environment management. A JavaScript version of the library is also in development, broadening the potential use cases and accessibility of the tool. Overall, Westworld aims to be a comprehensive and flexible platform for multi-agent simulation and optimization, suitable for academic, research, and practical applications in various fields.

https://github.com/TheoLvs/westworld

agent-behaviorsagent-communicationagentsalpha-releasecollectiblesepidemiologyexperimentalexportfleeinggifgrid-environmentsheuristicsimage-to-obstacle-conversionlogisticsmaze-generationmulti-agentmulti-agent-simulationnon-grid-environmentsobstaclespathfindingpygamepythonrandom-walkreinforcement-learningreplayretailrigid-body-systemsimulationsimulation-visualizationspatial-environmentsspawnerstable-baselinestriggersunity-ml-agentsvideovision-rangewanderingweb-integration