ValKyrie Q-Evac

Inspiration

In combat operations and major disaster responses, the "Golden Hour" dictates casualty survival rates. Traditional dispatch software relies on static shortest-path algorithms (Dijkstra/A*) that freeze or fail when dynamic threat vectors (artillery, SAM radars, moving hazards) shift in real time. We built ValKyrie Q-Evac to bring threat-aware, quantum-inspired route optimization to frontline emergency logistics.

What it does

ValKyrie Q-Evac is a tactical command center that optimizes medical drone and helicopter routing in high-threat environments: Triage-Aware Dispatch: Prioritizes casualty extractions based on medical urgency (Golden Hour countdowns) rather than basic proximity. QUBO Route Optimization: Formulates multi-objective pathing into a Quadratic Unconstrained Binary Optimization (QUBO) problem solved via a simulated-annealing sampler in ~90ms (with optional qiskit_optimization hardware mapping). Proportional Threat Avoidance: Dynamically calculates segment-to-threat geometric distance to penalize incursions while maintaining safe standoff perimeters. Fail-Safe Dashboard: A Next.js 14 tactical map interface featuring live telemetry polling, interactive casualty filter queues, and a zero-downtime offline fallback solver.

How we built it

Backend: Python + FastAPI handling high-speed JSON contracts via Pydantic (models.py). The core quantum engine (quantum_engine.py) builds haversine matrices, calculates segment-to-center threat vectors, and executes the QUBO sampler. Frontend: Next.js 14 (App Router), Tailwind CSS, Lucide React, and Leaflet rendering dark tactical tile overlays, pulsing danger zones, and real-time route polylines.

Challenges we ran into

Fixed threat penalty constants initially flagged almost every leg as high-risk due to scale relative to the Area of Operations (AO). We recalibrated the threat-exposure classification to normalize proportionally against each scenario's actual worst-case threat penalty, giving dispatchers realistic, high-fidelity telemetry.

Accomplishments that we're proud of

• Achieving clean ~90ms execution times for dynamic QUBO optimization against realistic multi-threat scenarios.
• Zero build errors across full Next.js TypeScript compilation and FastAPI backend testing.
• Implementing a graceful offline fallback solver (lib/offlineFallback.ts) to ensure zero-downtime interactivity during tactical operations.

What we learned

• How to balance conflicting multi-objective costs (distance vs. threat avoidance vs. triage priority) within a single QUBO cost surface matrix.
• Techniques for normalizing geographic threat vectors across scaled regional maps.

What's next for ValKyrie Q-Evac

• Integrating live ADS-B drone telemetry feeds for autonomous path execution.
• Expanding hardware execution from simulated annealing directly to IBM Quantum hardware QPUs via Qiskit Runtime.

Built With

• fastapi
• leaflet.js
• next.js
• python
• qiskit
• qubo
• tailwind.css
• typescript