About the Multi-Agent DSL Framework

This demo site showcases a research framework for building, orchestrating, and evaluating multi-agent systems using domain-specific languages (DSLs). It highlights the ATSLP scheduler, HCMPL cache, and CALK knowledge layer, along with reproducible performance dashboards and scenario-driven demos.

Artifacts

• Paper (PDF): View
• Code Repository: GitHub
• Reproducibility Guide: Steps

What to Explore

• Performance Dashboard: self-contained charts and metrics
• Scenario Demos: emergency, wildfire, smart grid, healthcare, traffic, warehouse, and USAR
• Video Analysis: client-side motion analytics and on-page reports

How to Reproduce

All experiments are reproducible with fixed seeds and version-pinned dependencies. See the reproducibility materials for exact steps. Reviewers can also run verify_academic_results.py to re-check results.