University of Toronto and Maersk Collaborate on Quantum-Driven Warehouse Layout Optimization
March 23, 2017
Maersk Explores Quantum Optimization in Warehouse Layout with University of Toronto
As early experimentation with quantum computing began to expand beyond routing and network theory in 2017, Maersk—the world’s largest container shipping company—turned its attention toward a foundational logistics problem: warehouse layout design.
In collaboration with the Creative Destruction Lab (CDL) at the University of Toronto’s Rotman School of Management, Maersk launched a feasibility study on March 23, 2017, exploring how quantum-enhanced optimization could improve warehouse layout efficiency across its inland distribution hubs.
The goal was to simulate optimal paths and item placements that minimized walking distances and picker congestion, especially in high-throughput environments. The results, while early-stage, demonstrated the potential of quantum-inspired algorithms in enhancing classical digital twin modeling and warehouse efficiency.
The Challenge: Optimizing Complex Warehouse Layouts
Warehouse layout design is a notoriously difficult optimization problem. It involves numerous variables, including:
SKU velocity (how fast a product moves)
Zoning constraints (e.g., cold storage, hazardous materials)
Picker congestion
Robotic pathfinding
Inventory turnover
Seasonal variation in order types
Classical warehouse layout algorithms often rely on heuristics or metaheuristic models such as ant colony optimization, genetic algorithms, and simulated annealing. These methods perform well but hit scalability limits when dealing with thousands of SKUs in dynamic layouts.
Quantum optimization, particularly inspired by models like quantum annealing or tensor networks, offers an alternative pathway by reducing solution spaces more efficiently.
The Maersk-CDL Research Collaboration
The study focused on a 500,000-square-foot Maersk warehouse located outside of Rotterdam, Netherlands, which handled both cross-dock operations and long-term inventory storage. The team used a hybrid model:
Classical digital twin simulation created a real-time representation of layout, SKU locations, robotic picker behavior, and human paths.
Quantum-inspired optimization algorithms were applied to specific layout submodules, including pick path clustering and zone rebalancing.
The core innovation came from quantum-inspired techniques developed by CDL's Quantum Machine Learning Lab, where researchers trained on D-Wave’s and IBM’s early-access quantum programming environments.
Rather than executing directly on quantum hardware (which remained noisy and small-scale in 2017), the team created QUBO (Quadratic Unconstrained Binary Optimization) formulations and deployed them using classical emulation techniques.
Key Results
According to the internal findings published by Maersk’s Innovation Division in April 2017:
Picker travel distances in simulated environments were reduced by up to 12.7% when using quantum-inspired layout suggestions versus classical heuristic-only models.
Bottleneck zones—areas of frequent congestion near high-demand SKUs—were reconfigured more effectively using quantum-inspired clustering.
Robotic forklifts exhibited smoother route transitions in simulations due to better sequencing of storage zones and travel loops.
Moreover, the research showed that even without direct access to fault-tolerant quantum hardware, quantum logic principles could still improve classical layout planning—what is now often referred to as “quantum-inspired optimization.”
Expert Commentary
Dr. Zachary McDonald, then a visiting fellow at the CDL’s quantum program, commented:
“Our focus wasn’t to run production problems on a quantum computer. It was to borrow from quantum mechanics and discrete mathematics to solve warehouse constraints more efficiently. In logistics, time is money—and 12% efficiency gains translate to millions.”
Maersk’s Chief Innovation Officer, Morten Engelstoft, added:
“Warehouse configuration has traditionally been a trial-and-error, labor-intensive process. Quantum-inspired planning offers a promising way to shortcut some of that iteration, especially in large hubs with tight turnover windows.”
Rising Interest in Quantum for Physical Logistics
While most quantum activity in 2017 centered on cryptography or financial modeling, logistics companies like Maersk began probing physical facility optimization as a novel use case. A key driver was the maturity of digital twins and real-time sensor integration, which enabled high-fidelity simulation environments.
The team also studied future scenarios where real-time layout reconfiguration could occur in robotic fulfillment centers—adjusting picking zones dynamically during peak periods like Black Friday.
This aligns with broader trends that emerged later in the decade, including:
Quantum digital twins
Adaptive warehouse automation
Autonomous robot fleet coordination
The Role of the Creative Destruction Lab
The CDL has long positioned itself as a bridge between quantum research and industry application. In 2016, it launched the world’s first seed-stage quantum machine learning accelerator, in collaboration with Google, Rigetti, and D-Wave.
Maersk’s participation marked one of the CDL’s earliest successful crossovers into supply chain and logistics. The project also inspired follow-up studies with Canadian Tire and DB Schenker in 2018.
Later, CDL participants like Xanadu and Zapata Computing would build commercial platforms supporting logistics clients in similar hybrid classical-quantum use cases.
Quantum Logistics: Laying the Groundwork
While the 2017 project did not result in immediate commercial rollouts, it laid the groundwork for:
Future Maersk investments in hybrid warehouse control software
Wider adoption of quantum-inspired modeling across the logistics sector
Interest from national logistics research labs in Europe and North America
By 2020, Maersk had begun testing early gate-model quantum software from IBM on container loading problems, showing how this 2017 project helped build foundational expertise.
Conclusion
The March 2017 collaboration between Maersk and the University of Toronto’s Creative Destruction Lab marked a critical turning point in applying quantum-inspired algorithms to real-world logistics problems. Focused on warehouse layout—a cornerstone of efficient global supply chains—the study showcased tangible gains in simulated picker efficiency and layout optimization.
While not yet running on quantum hardware, the methodology proved that even inspiration from quantum principles could yield measurable operational benefits. In doing so, Maersk positioned itself as a logistics leader prepared for a post-classical future, while CDL continued to demonstrate the role of academia in commercial quantum innovation.