Canadian National Railway Partners with Multiverse Computing for Quantum-Driven Rail Logistics Optimization

Introduction: Rail Logistics Meets Quantum Algorithms

As North America’s largest rail freight carrier by revenue, Canadian National Railway manages over 32,000 kilometers of track spanning Canada and the central United States. Its complex logistics network includes container ports, inland intermodal terminals, and hundreds of trains in motion daily—each requiring precise coordination of cargo, locomotives, and personnel.

Traditional optimization methods, even when enhanced with AI, often struggle with the combinatorial scale and dynamic uncertainty in railway logistics. In response, CN partnered with Multiverse Computing, a leading quantum software firm known for its Singularity platform—designed to run quantum algorithms on both quantum and classical systems. The initiative targeted quantum-inspired solutions for two high-impact use cases:

• Rail yard congestion minimization and train sequencing

• Rolling stock (wagon) allocation and routing optimization

Multiverse Computing’s Technology Stack

Multiverse Computing, based in San Sebastián, Spain, is best known for developing quantum algorithms that are executable on today’s classical hardware, a strategy that enables immediate commercial applications without waiting for fault-tolerant quantum machines.

The company’s flagship platform, Singularity, supports:

• D-Wave and IonQ quantum backends for annealing- and gate-based processing

• Quantum-inspired tensor network solvers

• Hybrid classical-quantum workflows that combine constraint solvers with quantum-assisted decision trees

For CN, this meant deploying optimization solvers that could handle:

• Large-scale combinatorial problems (e.g., train arrangement at congested depots)

• Operational constraints such as maintenance slots, weather impacts, and regulatory windows

• Real-time input streams from rail sensors, GPS, and fleet management software

Phase One: Yard Scheduling Pilot at Winnipeg Terminal

The initial pilot focused on one of CN’s busiest hubs: the Winnipeg rail yard, which sees hundreds of freight cars reclassified daily.

Key challenges:

• Scheduling trains to minimize yard congestion and maximize throughput

• Reducing delays in assembling outbound trains with multiple origins and destinations

• Accounting for rolling stock availability, crew shifts, and maintenance windows

Multiverse developed a quantum-inspired scheduling engine modeled after Quadratic Unconstrained Binary Optimization (QUBO) frameworks commonly used in D-Wave architectures. The algorithm processed:

• Real-time car arrival/departure data

• Yard track availability

• Time-window constraints on train dispatches

Results:

After six weeks of simulated testing and one week of shadow deployment:

• Yard congestion was reduced by 17%, as measured by average car dwell time

• Train sequencing improved throughput by 12%

• Operational conflicts (e.g., double allocations or track overlaps) dropped by 22%

Although still in simulation mode, CN noted these improvements could equate to millions in annual savings if rolled out at scale.

Rolling Stock Allocation and Routing: Quantum vs. Classical

The second part of the pilot evaluated quantum methods for rolling stock utilization—specifically how CN assigns available freight wagons across a network of customer routes and industrial sidings.

Challenges include:

• Matching wagon types (e.g., tankers, containers, grain hoppers) to specific cargo

• Preventing deadheading (empty returns) across long distances

• Balancing network load and maintenance schedules

Multiverse’s team modeled this as a constrained optimization problem, incorporating:

• Real-time fleet status from CN’s TMS and rail asset trackers

• Routing priorities based on customer SLAs

• Maintenance scheduling requirements

The solver employed a hybrid system:

• Simulated annealing as the core optimizer

• Quantum-enhanced probabilistic subroutines for non-linear constraint resolution

Comparative Findings:

• The hybrid quantum algorithm outperformed CN’s legacy scheduler by 8–11% on total distance traveled per wagon

• Empty car repositioning was reduced by 7.5%

• Fleet availability improved, with fewer idle wagons sitting unused between jobs

These results suggested not only cost savings but also enhanced carbon efficiency, as more effective wagon utilization meant fewer emissions per ton-kilometer shipped.

Industry Context: Rail Meets Quantum Globally

CN’s quantum initiative is part of a growing global trend:

• DB Cargo in Germany announced in 2021 it was exploring quantum scheduling for cross-border freight

• Indian Railways launched a program with TCS and QNu Labs for post-quantum encryption in control systems

• BNSF Railway has been studying optimization enhancements via machine learning and is reportedly evaluating quantum tech for use in intermodal hubs

The rail sector, with its heavy reliance on network design, scheduling, and equipment allocation, is particularly well suited to combinatorial optimization—a domain where quantum and quantum-inspired techniques shine.

Strategic Significance for CN

For CN, the pilot represented more than a technical trial:

• It signaled a shift from AI-only to hybrid quantum-classical architectures in logistics software

• Positioned CN as a first mover among North American carriers in adopting quantum optimization

• Opened the door for further collaborations with startups and universities through Canada's National Quantum Strategy

Executives at CN’s Innovation Group emphasized that this pilot was part of a broader 5-year roadmap to explore:

• AI + Quantum coordination for network rescheduling after weather disruptions

• Energy-efficient logistics, pairing electrified freight corridors with route optimizers

• Freight security, including post-quantum encryption for onboard communications

Integration and Next Steps

Following the December 2021 pilot, CN and Multiverse outlined a Phase Two:

• Expand yard scheduling to Toronto and Chicago terminals

• Begin integration of Singularity platform into CN’s centralized Rail Operation Dashboard

• Host workshops with operational staff to co-develop user-facing optimization tools

Longer term, CN aims to:

• Connect optimization engines with real-time control systems (yard switchers, signal boxes)

• Benchmark quantum optimization ROI vs. classical AI in monthly KPIs

• Participate in Canada’s Quantum Technologies Supercluster, helping shape national priorities

Challenges and Considerations

While the pilot showed promise, CN acknowledged limitations:

• Solver speed was a bottleneck in high-volume situations

• Integration with legacy IT systems (some decades old) required middleware development

• Organizational change management—train dispatchers and logistics planners required retraining to interpret quantum-derived route outputs

Multiverse responded by offering explainability modules, translating solver decisions into natural language rationales and actionable steps.

Conclusion: Laying Tracks for a Quantum Future

CN’s partnership with Multiverse Computing is a case study in the practical application of quantum optimization in complex freight logistics. By focusing on yard operations and wagon routing, two high-impact, high-cost areas, CN demonstrated that quantum-inspired techniques can yield measurable, operational results today—not just in theoretical projections.

This initiative sets the stage for broader industry adoption, where real-world gains in throughput, emissions, and cost-efficiency become the proving ground for quantum value.