Special | Report
Manufacturing Use Cases in IoT Era heading
AI, Quantum Computing Technology Accelerate Digital Transformation

ujitsu Limited is transforming itself from an IT company to a digital transformation company. Digital transformation effectively utilizes both conventional information technology (IT) and cutting-edge technologies to derive value out of data in order to provide innovative services and help transform business processes. Cutting-edge technologies such as artificial intelligence (AI), 5G communications, IoT, and quantum computing, accelerate digital transformation.

Fujitsu Limited offers AI solutions and services under the Zinrai: Fujitsu Human Centric AI brand; and in the quantum computing field, the company has developed the Quantum-Inspired Digital Annealer, an offering focused on solving combinatorial optimization problems at high speed. This article introduces the Digital Annealer and how it can benefit industries around the world.

Fig. 1: Fujitsu's transformed business model
Fig. 1: Fujitsu’s transformed business model
Fig. 2: Accelerating digital transformation
Fig. 2: Accelerating digital transformation

Takes Cue from Quantum Theory
AI has reached the practical application stage in several business areas. For example, AI can be used to: detect human behavior through video analysis, detect machine failures and defective products through image analysis, improve language translation accuracy, and improve the accuracy of sales forecasts for retail stores.

Does AI solve all business challenges? No. For example, even if the future sales of a retail store can be forecasted using AI, it is necessary to determine the types and quantities of products to be purchased and the procurement timing. Also, sales staff work shift planning in accordance with the sales forecast is needed. Quantum computing technology can help increase the effectiveness of AI by greatly increasing compute capacity.

In general, quantum computing often brings to mind the “quantum gate” systems for which Google and IBM have recently announced research results. However, for everyday use, significant advances in the technology are still needed. In contrast, advanced companies across all industries are looking to quantum computing technology in the form of “quantum annealing” for a specific sub-set of business challenges. A Canadian company is the first to offer a quantum computer; however, it is not practical for solving real world problems.

The development of Fujitsu’s Digital Annealer technology has been inspired by quantum phenomenon. The Digital Annealer architecture is specialized for solving combinatorial optimization problems at high speed using the annealing method. Digital Annealer is not a quantum computing technology, per se, but can provide more advantages today than quantum computers. For example, quantum computers have difficulty in maintaining the quantum state for a significant amount of time. Quantum computers also have difficulty in connecting Qubits together in order to process larger problems. In contrast, Digital Annealer operates in the same environment as ordinary classical computers, offers greater capacity for large-scale problems, and is a fully interconnected architecture. Hence, the Digital Annealer technology can solve real-world problems that both classical and quantum computers cannot.

Quantum Phenomenon for Everyday Uses
Combinatorial optimization problems require that the best combinations or sequences that meet given constraints are found from among a huge set of combinations. To explain combinatorial optimization in detail, the typical Traveling Salesman Problem example can be used. This problem aims to find the sequence of cities to visit at the least cost and the shortest route when a salesman visits each city only once. If a salesman must visit five cities, the number of possible combinations becomes the factorial of five; meaning 120 possible routes. When a salesman must visit 32 cities, there are 2.63×1035 possible routes. The number of possible combinations increases exponentially as the number of cities increases. Using a classical round robin approach, the calculation can surpass the capabilities of even modern supercomputers. The innovative Digital Annealer can calculate the optimal solution instantaneously.

Fig. 3: Quantum-Inspired Digital Annealer
Fig. 3: Quantum-Inspired Digital Annealer

Combinatorial optimization problems exist in various tasks across many industries. Combinatorial optimization problems can be found in the aforementioned example of retail purchase planning to maximize margins and employee work shift planning. Other examples include shipment planning optimization for distribution companies and the optimization of parts pickup routes in warehouses. The optimization of production planning and product design benefit the manufacturing industry. The fields of drug development and chemical materials research use combinatorial optimization for molecular structure similarity calculations and for finding stable structure molecules. In banking, the efficient calculation of combinatorial optimization problems can be used in the management of investment portfolios. Autonomous driving and the management of automated guided vehicles (AGV) make use of route optimization to alleviate traffic congestion.

Fig. 4: Combinatorial optimization problems
Fig. 4: Combinatorial optimization problems
Fig. 5: Real world applications of Digital Annealer
Fig. 5: Real world applications of Digital Annealer

The Fujitsu Digital Annealer has real world applications across a wide array of businesses. Since introduction in May 2018, Fujitsu has received an overwhelming response, with inquiries received from around the world. Through close collaboration among engineers, data scientists, and industry experts, Digital Annealer technology is providing benefits to research, development and business processes for many companies and organizations both in Japan and overseas.

Fujitsu is driving digital transformation with and for its customers with an incredibly wide array of technologies including AI, 5G, and IoT, as well as Digital Annealer.

About This Article:

The author, Shunsuke Tsuboi, is from the AI Business Division, Digital Business Development Unit, Fujitsu Limited.