- Paving the Way for Solving Combinational Optimization Problems in Mobility, Finance, Drug Discovery, and a Variety of Other Industrial Fields -

Main Points of Announcement
・In a wide range of industrial fields including advanced mobility services, smart factories, finance and drug discovery, which are the core of Society 5.0, there are various underlying combinational optimization problems.
・However, before we use Ising machines including quantum annealing machines that would provide solutions to the problems, there are many underlying difficulties.
・Therefore, in our research and development, we will develop a common software platform as an intermediate layer between Ising machine hardware and real problems.

On September 5, 2018, a group of six co-proposer organizations represented by Waseda University (Chief of R&D: Prof. Nozomu Togawa) and including Tokyo Institute of Technology, National Institute of Informatics, Research Organization of Information and Systems("NII"), National Institute of Advanced Industrial Science and Technology ("AIST"), Toyota Tsusho Corporation, and Fixstars Corporation, adopted "Development of Next-generation Computing Technologies," one of the research and development items in " Project for Innovative AI Chips and Next-generation Computing Technology Development," a project conducted by New Energy and Industrial Technology Development Organization ("NEDO"). The name of the adopted project is "Research and Development of a Common Software Platform for Ising Machines" ("this R&D"). The period of the project is from the 2018 fiscal year through the 2022 fiscal year. In addition, NEC Corporation will join in the project as a partner to conduct the project together with Waseda University.

Ising machines including quantum annealing machines are expected to provide a breakthrough for fast solutions to combinational optimization problems. In Japan, research and development of Ising machines, including quantum annealing machines and semiconductor-based Ising machines, has also been actively conducted, and various types of Ising machine hardware have been proposed and developed.

On the other hand, in various industrial fields including advanced mobility services, smart factories, finance and drug discovery, which are core fields of Society 5.0, there are various underlying combinational optimization problems, and there is a great gap between such real problems and Ising machine hardware aimed at resolving them. A key to making the quantum annealing machine a success is determining how to fill in the gap and make the quantum annealing machine easy to use.

To resolve the problems described above, in this R&D, we will conduct the research and development of a common software platform, including a group of middleware programs and common APIs as an intermediate layer between real issues and various Ising machine hardware. As a result, it is expected to make it possible to resolve various real problems, which have so far been difficult to apply the technology to, by using the common software platform on Ising machines.

Six organizations are aiming to jointly develop the common software platform for various Ising machines. In particular, the organizations will advance the development in close cooperation with "Project for AI Chips and Next-generation Computing Technology Development / Development of Next-generation Computing Technologies / Research and Development of Quantum Annealing Technologies Using Superconductive Parametron Devices" (Representative: NEC Corporation), which are NEDO's commissioned projects for developing new quantum annealing machine architectures.

To realize Society 5.0, social demands for high-performance computing as well as the progress of digitalization in various industrial fields, including advanced mobility services, smart factories, finance and drug discovery, have been growing rapidly. There are underlying “optimization problems,” especially “combinational optimization problems,” in industrial fields where Society 5.0 will be implemented. Therefore, it is said that obtaining fast optimum solutions to combinational optimization problems at high speed, even if they are difficult problems, will determine whether Society 5.0 succeeds or fails.

In this R&D, Ising machines, including quantum annealing machines, are expected to provide a breakthrough for fast solutions to combinational optimization problems. From sample data, it is said that combinational optimization problems would be solvable at high speed using the Ising machine. In addition, base technologies supporting the quantum annealing machine, for example, superconducting quantum bits and fast solutions to combinational optimization problems using the Ising model, are technologies that were proposed and demonstrated in Japan in 1990s.

Against this background, at present research and development of Ising machines, including quantum annealing machines and semiconductor-based Ising machines, is also actively in progress in Japan, and various types of Ising machine hardware have been proposed and developed. However, there is a great gap between real problems and Ising machine hardware that would provide solutions to such real problems. Therefore, how to fill in this “gap,” more specifically, how to build a “software platform” which is a common platform operating on various Ising machine architectures, as an intermediate layer between real problems and Ising machines, has become a key issue.

To resolve these problems, in this R&D we will research and develop a common software platform including a group of middleware programs and common APIs as an intermediate layer between real problems and Ising machine hardware. As a result, it will be possible to provide a common software platform that will operate not only on the Ising machines that have been developed so far in Japan and overseas, but also on various Ising machines to be developed in the future. It will also be possible to fill in the gap between real problems and Ising machine hardware, and to provide solutions to complicated and diverse real problems, on various Ising machines.

1.) Waseda University
・Development of basic algorithms and elemental technologies for a common software platform for Ising machines

2.) Tokyo Institute of Technology
・Development of basic quantum annealing theories for a common software platform for Ising machines

3.) National Institute of Informatics, Research Organization of Information and Systems
・Development of classical algorithms for the evaluation of a common software platform for Ising machines

4.) National Institute of Advanced Industrial Science and Technology (AIST)
・Development of security and material design applications

5.) Toyota Tsusho Corporation
・Study of the extraction and formulation of problems for the evaluation of a common software platform for Ising machines
・Development of next-generation mobility, logistics and supply chain applications

6.) Fixstars Corporation
・Development of libraries and APIs for the evaluation of a common software platform for Ising machines

7.) NEC Corporation
・Optimum design of hardware architectures for quantum annealing machines that will work with software

・WASEDA University (Headquarters: Shinjuku, Tokyo, Japan, President：Aiji Tanaka)
・Tokyo Institute of Technology (Headquarters: Meguro, Tokyo, Japan, President：Kazuya Masu)
・National Institute of Informatics, Research Organization of Information and Systems (Headquarters: Chiyoda Tokyo, Japan, Director General：Masaru Kitsuregawa)
・National Institute of Advanced Industrial Science and Technology (Headquarters Chiyoda Tokyo, Japan, President：Ryoji Chubachi)
・Toyota Tsusho Corporation(Head Office: Nagoya, Aichi, Japan, President & CEO: Ichiro Kashitani )
・Fixstars Corporation (Head Office: Shinagawa, Tokyo, Japan, CEO: Satoshi Miki)