Welcome to SMART

SMART is a simulator for multi-agent path finding, built on ARGoS. It is open-source, cross platform, and supports realistic physics-based simulation for evaluating MAPF algorithms with hundreds to thousands of agents.

Our goal is to develop SMART as a platform for MAPF research to bridge the gap between algorithmic planning and real-world deployment. For this purpose, SMART exposes APIs to integrate various MAPF planners and retrieve execution statistics.

Video Demo

Watch the SMART demonstration video on YouTube: https://www.youtube.com/watch?v=TX-oGSgM8VQ

How to Get It

Download

Download the latest release from GitHub Releases

Build from Source

Build on Linux - Build on Ubuntu/Linux
Installation - General installation guide

How to Use It

SMART provides multiple ways to explore and use the testbed:

Try the Web Demo: Experience SMART instantly at https://smart-mapf.github.io/demo/ - see Web Demo Tutorial for a full walkthrough
Quick Start: See Running SMART to launch your first simulation
Python API: Use the Core APIs for programmatic control
Planner Integration: Connect your MAPF algorithm via Planner Integration
Configuration: Customize simulations using Settings and Configuration

Check out our Examples for common use cases.

Tutorials

Examples - Getting started examples
Paper and citation - Research paper and benchmarks

Participate

Paper

More technical details are available in the SMART paper (arXiv:2503.04798).

Please cite this as:

@article{yan2025smart,
  title={Advancing MAPF towards the Real World: A Scalable Multi-Agent Realistic Testbed (SMART)},
  author={Yan, Jingtian and Li, Zhifei and Kang, William and Zheng, Kevin and Zhang, Yulun and Chen, Zhe and Zhang, Yue and Harabor, Daniel and Smith, Stephen F and Li, Jiaoyang},
  journal={arXiv preprint arXiv:2503.04798},
  year={2025}
}

Contribute

Who is Using SMART?

See Who is Using SMART? for projects and research using SMART.

What’s New

Initial release with ARGoS-based simulation
Support for 100+ agents
Action Dependency Graph execution monitoring
Python API for planner integration

FAQ

If you run into problems, check the FAQ - Frequently Asked Questions.

License

This project is released under the MIT License.