Operationalizing Quantum Optimization for Multi-Depot Logistics: A Hybrid service-oriented Approach

Abstract

This paper presents a hybrid quantum--classical microservices architecture for fleet management and vehicle routing in the NISQ era. We first introduce a reference architecture with 9 services that separates data acquisition, operational orchestration, and solving capabilities through classical and quantum routing services. We then formalize a baseline Multi-Depot Capacitated Vehicle Routing Problem (MD-CVRP) on a directed graph with explicit depot assignment, customer service uniqueness, flow conservation, capacity feasibility, and load propagation constraints. The paper also reviews state-of-the-practice approaches for solving MD-CVRP, from classical (exact and metaheuristic) methods to quantum-assisted QUBO/Ising workflows (QA and QAOA), and discusses their current practical limitations, including encoding growth, penalty calibration, feasibility repair, and quantum-hardware requirements. Based on this analysis, we outline a pragmatic research path on this problem: use strong classical baselines, apply quantum routines only to well-scoped subproblems, and report end-to-end performance with full accounting.