Westworld Simulation
Visit WebsiteGeneral Purpose
Open Source
A Python library designed to simulate and optimize environments with multiple interacting entities.
About
Westworld is a multi-agent simulation library inspired by Unity and Unity ML Agents, adapted for Python. It aims to simulate and optimize systems where multiple agents interact, making it suitable for applications in logistics, retail, and epidemiology. The library provides pre-coded spatial environments and facilitates communication between agents, with optimization achievable through heuristics and reinforcement learning. Note: Westworld is currently in an experimental, alpha-release stage and is under active development.
Features
- Easy creation of grid and non-grid environments.
- Support for various objects: Agents, Obstacles, Collectibles, Triggers.
- Ability to subclass objects for custom implementations.
- Spawner functionality for random object generation within environments.
- Basic rigid body system for all objects.
- Predefined agent behaviors: pathfinding, wandering, random walk, fleeing, vision range.
- Automatic maze generation capabilities.
- Layer integration to convert images into obstacles aligned with a grid.
- Sample simulations and agents for classic scenarios.
- Visualization, replay, and export options for simulations (GIF or video formats).
Tags
Simulation Library
Multi-Agent Systems
Python