Fleet distances have been reduced on realistic delivery benchmarks, with Accenture Labs India achieving reductions of seven thousand nine hundred and twenty-seven metres, equivalent to two point eight percent, on a twenty-five customer instance using Toshiba’s SQBM+ technology. This hybrid classical-quantum approach combines traditional optimisation techniques with quantum computing, demonstrating practical benefits in logistics whilst identifying current limits of hardware at solving these complex problems; QAOA reliably solves subproblems up to four customers requiring sixteen qubits. Researchers pioneered a new method combining traditional computer programming with early forms of quantum computing to optimise delivery routes in complex city environments, specifically modelling London traffic patterns.
This hybrid approach demonstrably reduced travel distances in testing, proving its potential even given current limitations in quantum technology. Key analysis revealed that tight scheduling constraints present the biggest obstacle to further improvements using existing quantum computers, rather than optimising individual route segments. Optimising delivery routes in congested cities remains a vital challenge, with complex constraints limiting traditional methods of finding efficient solutions.
Researchers developed a hybrid approach combining standard computer programming, using techniques like Mixed Integer Linear Programming to establish feasible starting points, akin to carefully packing boxes into a truck, with early forms of quantum computing. This system tackles portions of the problem designed specifically for Quadratic Unconstrained Binary Optimisation (QUBO), which can be visualised as a puzzle where switching pieces “on” or “off” unlocks an optimal answer; this is then used to refine potential routes.
Testing on models of London traffic revealed that strict delivery deadlines present the biggest hurdle to further improvements with current quantum computers, how tightly scheduled deliveries are similar to arranging meetings within limited time slots.
SQBM+ delivers initial gains but schedule constraints limit broader quantum vehicle routing
Fleet distance reductions of 2.80%, equivalent to seven thousand nine hundred and twenty-seven metres, were achieved on a twenty-five customer instance using Toshiba’s SQBM+ technology compared to traditional methods; this represents a sharp leap in optimisation potential previously unattainable at such scale. Current quantum hardware reliably solves constraint satisfaction problems only within limited parameters, specifically up to four customers requiring sixteen qubits before encountering intractable complexity, highlighting existing boundaries for practical application.
Diagnostic results revealed that assignment-level time window tightness is the primary obstacle preventing further scalability improvements with current systems, rather than optimising individual route segments; these insights guide future work towards addressing core limitations instead of refining already efficient processes.
Experiments utilising QAOA across multiple hardware platforms confirmed reliable constraint satisfaction only for subproblems involving up to four customers requiring sixteen qubits maximum, establishing clear limits for current quantum processing capabilities. Detailed analysis revealed that the strictness of assignment-level time windows hinders scalability improvements within existing systems more than optimising individual route segments; these findings position hybrid classical, quantum architectures as a viable near-term solution for logistics challenges while also stressing where further advancements in quantum hardware are essential before widespread practical application becomes feasible.
Fleet distance reduction balances against hardware scalability challenges
The team’s hybrid approach demonstrably reduces fleet distances on models of London traffic, an important step towards practical quantum logistics, but current hardware imposes stark limitations. Reliable results were only obtained with problems involving four customers or fewer; scaling beyond this point quickly becomes intractable due to qubit limitations and circuit complexity.
Despite current hardware limiting reliable quantum solutions to subproblems involving up to four customers, this work demonstrates a viable path toward practical quantum logistics. Identifying assignment-level time window tightness as the primary obstacle allows focused development efforts and justifies continued investigation into decomposition strategies that avoid variable growth and unstable constraint enforcement in larger problems. SQBM+ achieved verified fleet distance reductions of 2.80% (7,927m) for the twenty five customer instance and 0.22% (1,186m) for the fifty five customer instance by combining classical techniques with emerging quantum approaches. The team will next focus on overcoming these time window restrictions to unlock further improvements; deliberately augmenting classical methods with quantum computation provides measurable benefits in logistics despite current hardware limitations and offers evidence for the viability of such systems in near-term applications. Identifying assignment-level time window tightness as the dominant scalability bottleneck opens new avenues for work focused on algorithms better suited to handling strict scheduling constraints rather than simply optimising individual routes.
This research successfully combined classical optimization techniques with a quantum approach using Toshiba SQBM+ and Rigetti/IQM processors to address complex vehicle routing problems based on London traffic data. The hybrid method achieved verified reductions in fleet distance, 2.80% (7,927m) for instances with 25 customers and 0.22% (1,186m) for those with 55 customers, demonstrating the potential of this combination. However, experiments revealed that current quantum hardware reliably solves only subproblems involving up to four customers due to limitations in constraint satisfaction; identifying time-window tightness as the primary obstacle will guide future work focused on improving scalability.
👉 More information
🗞 Hybrid Classical-Quantum Optimization for CVRPTW: A Systems Framework with Provable Feasibility and Empirical Hardware Limits
✍️ Shubhashis Sengupta (Affiliation: Accenture Labs India {syamasundar.gopasana); Syamasundar Gopasana, Swatantra Singh and Rohan Joydhar
🧠 ArXiv: https://arxiv.org/abs/2611.15343




See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
