NTT DOCOMO Deploys Second D-Wave Quantum Application

NTT DOCOMO, Japan’s largest mobile operator, has implemented a second D-Wave-powered quantum application to optimize how its network groups geographic areas. According to DOCOMO, this application reduced peak location registration signals by 65.3% and paging signals by 7.0% in a test network scenario. This is a live deployment, though the reported figures are based on a single representative problem.

Application details

Per the August 18 press release, the application optimizes what DOCOMO refers to as tracking area lists, or TA-Lists. These lists group sections of a mobile network to balance signaling load with connectivity. DOCOMO’s initial D-Wave application, already in production, focused on reducing paging signals within individual tracking areas. This second application operates across multiple tracking areas simultaneously, managing the often-conflicting demands of location registration signals and paging signals.

In the specific problem described by the companies, the system covered 333 base stations, three TA-Lists, and nine tracking areas. The optimization process was completed in approximately five minutes. DOCOMO states that these results inform its network planning and daily operations. The reported 65.3% and 7.0% reductions pertain to this specific daily peak scenario and not the entire national network.

The systems leverage D-Wave’s hybrid annealing technology, which combines classical computing with quantum annealing hardware. D-Wave, listed on Nasdaq as QBTS, positions itself as the sole company offering both annealing and gate-model quantum systems.

Company Statements

D-Wave CEO Alan Baratz presented the deployment as a signal to other telecommunications operators. Hiroshi Kawakami, Senior Manager of DOCOMO’s Data Platform Department, commented on the technical outcomes. As Kawakami stated:

“With our second production application powered by D-Wave’s hybrid quantum technology, we are applying quantum optimization to a broader and more complex area of network operations. The application balances two competing network requirements, significantly reducing location registration signaling while also lowering paging signals.”

Notably absent from the release are figures regarding DOCOMO’s financial commitment to D-Wave, a timeline for wider implementation across its network of over 93 million subscriptions, or a comparison demonstrating how a classical solver would perform on the same 333-base-station problem. 

Though a five-minute runtime and percentage reduction in a test case confirm the application’s functionality, they don’t establish that quantum annealing surpasses conventional optimization methods in this context, as no head-to-head benchmark is provided.