QADA: Quantum Fleet Optimizer
A hybrid quantum-classical vehicle routing system that solves real-world delivery logistics using two-level QAOA optimization — running actual circuits on IBM's 156-qubit Heron quantum processor.
Qubits
156
QPU Runs
25+
Architecture
Two-Level QAOA
Hardware
IBM Heron
The Problem
The Vehicle Routing Problem (VRP) is NP-hard — as fleet sizes and delivery points grow, classical solvers hit computational walls. Traditional approaches like brute-force enumeration or greedy nearest-neighbor heuristics either don't scale or produce sub-optimal routes that waste fuel, time, and money.
Two-Level QAOA Architecture
Inner Level — Per-Zone TSP: For each delivery zone, a QUBO is built from the orders-only distance matrix. QAOA finds the optimal Hamiltonian path through orders within the zone. Forward and reverse traversal costs are cached for the outer level.
Outer Level — Inter-Zone Routing: A meta-QUBO encodes K zones as 2K virtual nodes (forward/reverse ports) with 2K² binary variables. QAOA on this meta-QUBO returns optimal zone ordering and per-zone traversal direction. No brute-force, no 2-opt repair, no greedy fallback.
Variational Training: SPSA optimizer on a host-side Aer Estimator finds best-iterate parameters across the trajectory. Final shot-based execution runs on the real QPU.
Interface
