Graduate-level Research in Industrial Projects for Students (GRIPS) – Sendai 2019

June 17 - August 9, 2019

Sponsors and Projects

The sponsors and projects for 2019 include:

Project 1: TOYOTA Motor Corporation: Mobility service for university campus:

Project 2: TOYOTA Motor Corporation: Mobility service for hospital guests

Project 3: FUJITSU Laboratories Ltd.: Resolving real-world issues by “Digital Annealer”

Project 4: NEC Corporation: Combinatorial optimization using quantum annealing: Search for proper choices of solvers and evaluation of solutions on combinatorial optimization problems


Note: At this time we can only accept applications for the first Toyota project: Mobility service for university campus


Project 1 and 2: TOYOTA Motor Corporation: Design for the next generation mobility service in suburban areas


Toyota will lead the way to the future of mobility, enriching lives in the world with the safest and most responsible ways of moving people. In the near future, cars are expected to connect with people and communities and take on new roles as part of the social infrastructure. New areas, such as AI, autonomous driving, robotics and connected cars, are becoming especially important. Toyota aims to reach the ultimate goal of sustainable mobility, creating a mobile future society with full of smiles.


Rapid suburban sprawl in metropolitan areas in Japan has led to the rapid development of peripheral satellite cities. High-speed railway networks and motorways have significantly contributed to improving the connection between established metropolitan areas and these new peripheral cities. But at the same time, due to rapid population growth, sprawled suburban areas with small-scale residential developments were built without schematic road network planning.

As a result, key functions of the city for the citizen’s lives, such as commerce, administration, and schools, are diffused to suburban area, on the premise of using private cars. But such developments are not sustainable or desirable as the population ages and as energy conservation becomes more important. In principle, a higher quality of life can be achieved if key city functions are aggregated in city centers around the main railway stations. This requires better and more efficient public transportation networks in conjunction with small personal mobility.

Mobility-as-a-Service (MaaS) describes an integration of different transportation modes towards mobility solutions that are consumed as a service. Mobility innovation, often called as CASE (C: Connected, A: Autonomous, S: Sharing, E: Electric) could be fully or partially implemented as MaaS. Wider range of demonstration programs and commercial based services are being provided to improve safety, efficiency and convenience of transportation.


There are many studies on MaaS as implementation of CASE, particularly for metropolitan areas where other multiple transportation modes such as rails, metro and buses are available. But at the same time, other areas, where private cars are dominant transportation mode, are not thoroughly investigated. Mobility services for non-metropolitan areas could be different from that for metropolitan areas, due to differences in population density/distribution, transportation demand, and public transportation network.

As an example, Tsukuba city, a city with a population of 237,000 people, and located 60 km northeast of Tokyo, has a private car dependent transportation system. In this project, we organize two groups of students as Group-1 and Group-2 in which the students will design MaaS for practical cases, in particular for university campus and hospital guests, through analyses on existing person-trip and other related data sets.

Project 1: Mobility service for university campus:

University of Tsukuba has a relatively large campus area and a large number of students. The present major transportation modes are buses, bicycles and walks. Students are expected to build a practical planning model to optimize the MaaS for University of Tsukuba campus.

Project 2: Mobility service for hospital guests:

University of Tsukuba Hospital is an advanced treatment hospital in this region. There is a strong need for transportation service for hospital visitors, since the hospital is a few kilometers away from a major transportation hub.Students are expected to build an efficient MaaS planning to optimize the mobility service for hospital guests considering characteristics of several types of patients.

The problems setups and formulations may be different for these targets and the two groups will work independently during the GRIPS-Sendai program. However, exchanging ideas and discussions between these two groups will be highly encouraged and we expect some universal idea for MaaS in the future society. Although there already exist several approaches for this kind of problems, students are expected to find totally new mathematical approaches, descriptors, formulations, solvers, visualizations, operation plans, etc. to be fit for our future society with highly sophisticated mobilities. This project includes sight visits to a mobility exhibition center in Tokyo and University of Tsukuba.


 We welcome applications from motivated team-players who have knowledge and practical skills of one or more of the followings:

  • Mathematical statistics
  • Optimization
  • Operations research
  • Programing language (Python, C, or MATLAB)


Project 3: FUJITSU Laboratories Ltd.: Resolving real-world issues by “Digital Annealer”


Fujitsu Laboratories was established in 1968 as an organization independent from Fujitsu for conducting the world’s top-level technological development in a free atmosphere for researchers. Fujitsu is the leading Japanese information and communication technology (ICT) company, offering a full range of technology products, solutions and services. We have produced a lot of innovative results through IT and have been leading the world by providing new values to the people, society and businesses. Fujitsu is the world’s 7th largest IT services provider and No.1 in Japan.


Digital Annealer (DA) a new technology to solve large-scale combinatorial optimization problems instantly. It uses a digital circuit design inspired by quantum phenomena and can solve problems that are tough for classical computers to deal with. Many real-world social issues such as environmental problems, can be regarded as combinatorial optimization problems. The objective of this project is to solve a serious societal issue using DA.

Quantum computing technologies are categorized into two types. One is quantum gate computer, and the other is Ising machine [1, 2]. Quantum gate computers are for universal computing, whereasIsing machines are specialized in searching for solutions of combinatorial optimization problems.

Moreover, there are two types of Ising machines. One is the quantum annealing machine and the other is the simulated annealing machine. The quantum annealing machine searches solutions by using quantum bits which are made of quantum devices such as a superconducting circuit. The simulated annealing machine uses a digital circuit. DA is a type of simulated annealing machines with a new digital circuit architecture that is designed to solve combinatorial optimization problems efficiently. The number of bits (node) of an annealing machine is closely related to the number of combination parameters. Annealing machines with large number of bits (nodes) can solve large-scale combinatorial optimization problems.

Many of the world’s social issues can be treated as combinatorial optimization problems that cannot be easily solved by conventional computers when the problem size increases. The applicable fields are chemistry, finance, transportation etc. DA uses the MCMC (Markov chain Monte Carlo) method [3] and the SA (Simulated Annealing) method [4, 5] to solve an Ising model [6-9] to which a combinatorial optimization problem is converted. This conversion technique is very important to solve problems at high speed.

The general procedure to solve problems by DA is shown as follows.

1)  Problem description

2)  Formulation to combinatorial optimization problem

3)  Re-formulation to Ising model

4)  Conversion into QUBO (Quadratic unconstrained binary optimization)

5)  Calculation of optimal solution by DA

Fujitsu has launched a DA cloud service in May 2018. The first generation has 1024 bits, all of which are fully connected with 16-bit precision. Although some problems can be handled within this bit size, many of the real-world problems need a larger bit scale. We have developed the second generation service of DA using DAU (Digital Annealer Unit) which has an 8K bit scale. By making both hardware and software enhancements, DA can deal with problems on a scale of 100K bits.


The main aim of this project is to find concrete formulations and efficient algorithms fitted to DA and ultimately find optimal solutions for the combinatorial optimization problems. Students will obtain an experience of thinking about solving social issues in the real world, and observe that many of real-world problems are regarded as combinatorial optimization problems. The project focuses on flow optimization problems among a wide variety of combinatorial optimization problems. Examples of flow optimization problems are as follows:

1) Route optimization for avoiding traffic congestion:

Reduce overall travel time by assigning different routes throughout the city or the entire factory.

2) Optimization of work flow line in a factory to improve productivity:

Up to a 45% reduction in moving distances in a warehouse was achieved by Fujitsu IT Products

Students can select one of these examples and try to find new formulations, or find a new societal problem. Students are expected to look into the problem scale that can be calculated with DA and the combination scales of real problems. It is also expected that the students try to devise a method of obtaining an accurate calculation result. Fujitsu will provide user accounts and computational environments of the new DA machine to the member of this project during the period of GRIPS-Sendai 2019.


Programming (mandatory: Python, optional: C)


Project 4: NEC Corporation: Combinatorial optimization using quantum annealing: Search for proper choices of solvers and evaluation of solutions on combinatorial optimization problems


NEC was founded in 1899 and is now focusing on developing solutions for the society that will help resolve many of the issues the world is facing and lead to the creation of a brighter and more prosperous society. Through co-creation initiatives with many different stakeholders, including customers and business partners, citizens and government agencies, and international institutions, we are actively devising new business models to create social value by harnessing our extensive ICT (Information and Communication Technology) assets.


Artificial intelligence (AI) is increasingly being used to solve various real-world problems. One method for solving problems is selecting the best combination from an infinite number of choices generated through machine learning.

These kinds of complex combinational optimization problems take even supercomputers an enormous amount of time to solve, making it impossible to obtain accurate answers within the required time. Business, however, requires making the best decisions within at most a few tens of minutes.

As societies mature, their problems also increase in complexity, leading to an ever-increasing number of factors to consider in solving problems. In the future, the evolution of computer technologies capable of solving complex problems will lead to significant social changes across many fronts: time savings, energy and cost reduction, proper manpower allocation, labor shortage resolution, industrial optimization, and value creation.

Among quantum computers, quantum annealing machines, which use weights among qubits (Quantum bits) in performing computations, has received wide attention as a tool for solving complex combinational optimization problems within a short period of time. In 1999, NEC succeeded in demonstrating solid-state qubit operation in the quantum superposition state for the first time through a superconducting device. As it continued its research on ways to control the superposition state, NEC at some point shifted to quantum annealing as the main focus for generating a suite of outcomes from its research efforts.

Project and Expectations:

Recent advent of quantum annealers and digital annealers has brought a breakthrough in solving combinatorial optimization problems fast and accurately. However, size of target problems, for example, obstructs us in directly applying these solvers to the problems, and conventional approaches often outperform the new solvers.

Students are subjected to propose proper choices (or combinations) of the annealers and/or solvers, for Employee Shift Scheduling as an example, which give the most accurate solution in the shortest time. Students are also subjected to develop formulation of corresponding Ising models(*1) and estimate the applicability of each solver. NEC will provide datasets for these evaluations.

1)            The quantum and classical annealers provided by D-wave systems Inc.(*2),

2)            Solvers based on Xeon servers,

3)            Solvers based on GPGPU servers,

4)            Solvers based on NEC’s Vector machine “SX-Aurora TSUBASA” servers(*3),

and their combinations are also available.

(*1)The Ising model, named after the physicist Ernst Ising, is a mathematical model of ferromagnetism in statistical mechanics. The model consists of discrete variables that represent magnetic dipole moments of atomic spins that can be in one of two states (+1 or −1). The spins are arranged in a graph, usually a lattice, allowing each spin to interact with its neighbors. Quantum annealers and digital annealers try to find the minimum energy state of this model, which corresponds to the optimal combination of the target problem.

(*2) Students can use the classical annealer provided by D-wave systems Inc. to solve the problems Then their Ising models will be executed by industrial mentors on the D-wave quantum annealer.



Programming skills (Python, C, or MATLAB)

Recommended Readings and References

[1]          Professor Hidethosh Nishimori’s website on quantum annealing.

[2]          D-Wave’s website.

[3]          M. W. Johnson, et al., “Quantum annealing with manufactured spins,” Nature 473, 194 (2011).

[4]          M. Aramon, et al., “Physics-Inspired Optimization for Quadratic Unconstrained Problems Using a Digital Annealer,” arXiv:1806.08815.

(*4) You’ll find materials on some applications of quantum annealing on this website.