NTT DOCOMO Expands Production Use of D-Wave Quantum Technology

Japan's Largest Mobile Carrier Faces a Growing Signaling Challenge

NTT DOCOMO (DOCOMO), Japan's largest mobile operator, serves more than 93 million subscriptions across a national network of roughly 250,000 base stations. Every phone, smartwatch, and connected device relies on two types of background signals to remain reachable. Devices send location-registration signals to update the network as they move into new areas, while the network sends paging signals to locate a device when a call or message arrives.

Like many telecommunications operators, DOCOMO manages these signals by dividing its network into tracking areas (TAs), which are grouped into larger tracking-area lists (TA-Lists). The network sends paging signals within individual TAs, while a device sends a location-registration signal when it moves from one TA-List to another. 

Reducing both signal types is a longstanding challenge for DOCOMO and the broader telecommunications industry because changes that lower one can increase the other. Smaller TA-Lists trigger more frequent location-registration signals, while larger TA-Lists require paging signals to be sent across more base stations when a call or message arrives. 

At DOCOMO's scale, the underlying optimization problem is enormous. For example, a network segment with 333 base stations grouped into three TA-Lists and nine TAs has 28991 possible configurations, vastly more than the number of atoms in the universe. The model must also account for operational constraints, such as avoiding registration boundaries along high-traffic areas like rail lines and highways. Multiplied across DOCOMO's full national footprint, the search space is far beyond what classical solvers can address on a production-relevant timeline.

Building on Quantum Success in Production

DOCOMO had already demonstrated the value of quantum optimization in its network. Since 2024, the carrier has been using D-Wave's constrained quadratic model (CQM) hybrid solver in a production application to optimize paging-signal traffic within individual TAs. Using mobile traffic data from three regions, the application reorganized 270 base stations into 21 TA-Lists. The problem took DOCOMO’s classical computing methods 27 hours to solve, compared with just 40 seconds using D-Wave’s platform. Once deployed in production, the application reduced peak-period paging signals by roughly 15%, giving DOCOMO’s network an estimated 20% more capacity for connected devices.

That result gave DOCOMO the technical foundation to apply quantum computing to a harder problem: optimizing across TA-Lists and TAs together to reduce paging and location-registration signals simultaneously. 

Optimizing Two Competing Signals at Once

DOCOMO aggregated real device-movement data between base stations across its network and formulated a multi-objective optimization problem. The application reassigns base stations to TA-Lists and TAs to minimize both paging and location-registration signals, balancing the trade-off between the two. It also avoids placing registration boundaries along major transit routes, where fast-moving devices would otherwise trigger bursts of unnecessary signaling, increasing the burden on base stations.

From a Five-Minute Solve to a Production Application

DOCOMO addressed the problem using D-Wave's CQM hybrid solver, which combines classical computing with the Advantage2™ annealing quantum computer, through the Leap™ quantum cloud service. For a segment of 333 base stations, three TA-Lists, and nine TAs, the solver completed the optimization in approximately five minutes. DOCOMO’s previous classical methods could not solve this problem within the time required for production use. 

DOCOMO also built a map-based visualization tool that converts the application’s output into a base-station reassignment plan. Regional branch offices can review the plan and estimated signal reductions before approving deployment. The application has now been rolled out across DOCOMO’s live national network.

The graph shows the reduction of location-registration signals after the implementation of the quantum network optimization application.

Across the 333 base station segment, the application reduced the daily peak volume of location-registration signals by 65.3% and paging signals by 7%, solving a challenge engineers previously thought impossible to address. Together, the reductions lower the network’s control-plane load and free up capacity to support continued growth in connected devices.

Hear from NTT DOCOMO at Qubits Asia

Join us at Qubits Asia in Seoul on October 28 to hear Shuichi Fukuda, a researcher at NTT DOCOMO, discuss the company’s production use of quantum computing for network optimization. This full-day quantum computing user conference will bring together D-Wave executives, customers, researchers, and industry leaders to share lessons from real-world implementations, exchange best practices and explore opportunities to accelerate quantum computing adoption across APAC.

Register