A discrete-event simulation platform for hospital patient flow, built with SimPy and Streamlit.
FAER simulates A&E → Resus → disposition pathways over 1-36 hour horizons, helping healthcare teams understand:
- Where congestion actually originates
- Which capacity investments have system-wide effects
- What fails first under surge conditions
- Full A&E Pathway: Triage → ED Bays → Theatre/ITU/Ward → Discharge
- Priority Queuing: P1-P4 triage levels with clinical priority handling
- Multi-Stream Arrivals: Ambulance, helicopter (HEMS), and walk-in patients
- Diagnostics: CT scanner, X-ray, and Bloods with realistic ED bay blocking
- Aeromedical Evacuation: HEMS and fixed-wing transfers to specialist centres
- Interactive Schematic: Real-time React/SVG visualisation of patient flow
- Clinical Insights: Heuristic-based agent that detects risk patterns (NHS thresholds)
- Scenario Comparison: Run multiple configurations and compare outcomes
# Clone the repository
git clone <repository-url>
cd Pj-FAER
# Install in development mode
pip install -e ".[dev]"# Run the Streamlit app
streamlit run app/Home.py
# Run tests
pytest tests/ -v
# Run with coverage
pytest tests/ --cov=faer --cov-report=htmlPj-FAER/
├── src/faer/ # Core simulation library
│ ├── core/ # Scenario config, arrivals, distributions
│ ├── model/ # SimPy patient processes
│ ├── results/ # Event logging and metrics
│ ├── experiment/ # Replication runner, CI analysis
│ └── agents/ # Clinical shadow agents for insights
├── app/ # Streamlit web UI
│ ├── pages/ # App pages (Arrivals, Resources, Run, Results, etc.)
│ └── components/ # React schematic component
├── tests/ # Pytest test suite
├── docs/ # Documentation and PRDs
└── config/ # Configuration files
| Page | Description |
|---|---|
| Arrivals | Configure arrival rates (simple, 24-hour profile, or detailed) |
| Resources | Set capacity for ED bays, theatre, ITU, ward, diagnostics |
| Schematic | Interactive patient flow diagram |
| Aeromed | Configure aeromedical evacuation settings |
| Run | Execute simulation with progress tracking |
| Results | KPIs, utilisation metrics, and flow schematic |
| Compare | Side-by-side scenario comparison |
| Sensitivity | Parameter sensitivity analysis |
| Insights | Clinical shadow agent risk detection |
- SimPy 4.1+ - Discrete-event simulation engine
- sim-tools - Tom Monks' DES utilities
- Streamlit 1.30+ - Web UI
- React/TypeScript - Interactive schematic component
- NumPy, Pandas, SciPy - Numerical computing
- Plotly - Visualisation
- CLAUDE.md - Development instructions and coding standards
- docs/PRD_AGENT_INTEGRATION.md - Clinical agent integration specification
This project uses patterns and concepts from:
- sim-tools by Monks, T., Heather, A., Harper, A. (2025) - GitHub - MIT License
- STARS Project by pythonhealthdatascience - MIT License
- simpy-streamlit-tutorial by health-data-science-OR - MIT License
MIT License - see LICENSE file for details.