The Next Generation of Energy Distribution System

Social Cooperation Research Departments

Energy is essential for the modern society. The spread of renewable energies is a key factor for achieving a sustainable world. In order to maximize the value of the diversified energy resources, it is important and highly required to develop new energy systems. In this research department, we utilize data science and applied mathematics to solve a variety of problems related to energy systems.

Projects
Works
Members
Access

Projects

The issues that we plan to work on from April 2021.

DERMS

The output of renewable energies, such as photovoltaic power generation, is unstable. To build a stable system we need to combine these devices with storage batteries and fuel cells. DERMS (Distributed Energy Resource Management Systems) is a mechanism to control such systems. In this research department, we develop such systems using ABM (Agent-based modeling) and mathematical optimization methodologies.

EV Fleet Management

It is difficult to optimize the charging schedule for a large number of electric vehicles (EVs). In addition, you might have to optimize not only the schedule but also the charging costs. These problems are known as a multi-objective optimization problem. We think this would be an interesting application of evolutionary computation such as differential evolution.

JEPX price forecasting

The price of electricity traded on the market (JEPX) occasionally deviates from the daily price variation. This stems from a variety of factors and is one of the hot research topics in political economy field. In this research department, we plan to apply agent-based modeling (ABM) to JEPX price forecasting in order to integrate the data science-based approach with the methodologies of political economy.

Works

The followings are results from bioinformatics research.

Drug repositioning

Drug repositioning (or repurposing) is the methodologies that extend the indication of the already approved drug to the other disease. We identified several drug candidates for Alzheimer's disease by using the method of network embedding with the deep neural network (autoencoder).

Alzheimer's Research & Therapy volume 13, Article number: 92 (2021)
A new network centrality

The importance of a node in the complex networks can be quantified by centrality measures. There are many types of centralities. Betweenness centrality is one of the most frequently used one, and we have made the variant of it called node-limited betweenness centrality (nlBC). We applied the centrality to the protein-protein interaction networks and indicated that it could detect the nodes which were neglected with the commonly used centralities.

BMC Systems Biology volume 6, Article number: 124 (2012)
Predicting the response to anti-cancer drugs.

We build the machine learning model that could predict the potential anti-cancer drug responder for the colorectal cancer patients. We used Random Forests algorithm and we could detect the gene that might related to the anti-cancer drug response mechanisms.

British Journal of Cancer volume 106, pages126–132 (2012)

Members

This research department is supported by AAKEL TECHNOLOGIES INC.

Professor
Masakazu SUGIYAMA
Sugiyama Lab.
Project Associate Professor
Shingo TSUJI
Data Science, Bioinformatics, Applied Mathematics, Python. Linkedin

We are collaborating with Genome Science & Medicine Division.

Access

Location
CCR-B508
Research Center for Advanced Science and Thechnology, The University of Tokyo
E-mail
tsuji@genome.rcast.u-tokyo.ac.jp