Research Groups Research

Flow and Physics イラスト

Group 01.

Flow and Physics

Model the “flow” of people and goods, and contribute to a mobility society in which complex transportation networks flow smoothly.

Model the “flow” of people and goods, and contribute to a mobility society in which complex transportation networks flow smoothly.

The “flow” created by self-driving cars and people is the blood that circulates through the body of the mobility society, so to speak. To prevent this flow from becoming stagnant, the Flow and Physics Group is developing mathematical models of the flow, as well as computational methods to guide the flow to its optimum level. Moreover, by using the concept of tensor networks, we aim to construct a revolutionary method that will make calculations overwhelmingly more efficient and eventually be able to handle the traffic flow of an entire city.

Research Outline

Our group is working on the development of new computational methods for various flows, including traffic flow, human flow, and even the movement of cells in living organisms. Our research procedure involves first examining the contents to be computed and then modeling them by incorporating various perspectives, including discrete integrable systems (Box–ball systems). Moreover, the created flow model is put into an algorithm and a computer simulation is performed. By developing a variety of methods to visualize flows, we hope to find a completely new perspective that captures the essence of the mobility society.
Furthermore, if we envision a future society in which self-driving cars are implemented, then it will be necessary to view traffic flow as a network and design an ideal network in which congestion does not occur. Although computing such a complex and huge network as the actual traffic network is challenging at this stage, our group is expecting to solve this problem by using the concept of tensor networks. Data with various indices such as position, time, and velocity are called tensors, and a tensor network is a simpler representation of these tensors in the form of a graph. By handling huge volumes of data in the form of tensor networks, the amount of computation can be greatly compressed.
Our goal is to deepen our knowledge by working on flow models and tensor networks, respectively, and to build a network flow model that can be applied in the future in a mobility society.

Participating Members

Satoshi TsujimotoGraduate School of Informatics, Kyoto University
Ryosuke KojimaGraduate School of Medicine, Kyoto University
Kazuki MaedaFaculty of Informatics, The University of Fukuchiyama
Zanlungo FrancescoInternational Professional University of Technology in Osaka
Masashi IwasakiFaculty of Life and Environmental Sciences, Kyoto Prefectural University
Akiko FukudaCollege of Systems Engineering and Science, Shibaura Institute of Technology
Kenji HaradaGraduate School of Informatics, Kyoto University
Naoki KawashimaThe Institute for Solid State Physics, The University of Tokyo
Tsuyoshi OkuboGraduate School of Science, The University of Tokyo
Daisuke TakahashiSchool of Fundamental Science and Engineering, Waseda University
Katsuki KobayashiGraduate School of Informatics, Kyoto University
Julien GaboriaudGraduate School of Informatics, Kyoto University
Atsushi MaenoGraduate School of Informatics, Kyoto University
Kazuya OkamotoSchool of Fundamental Science and Engineering, Waseda University
Motokazu YamashitaGraduate School of Science, Kyoto University