|11 - 13 November 2020, Phuket, Thailand|
RoboCup soccer is an international project of robotics and artificial intelligence, which has an ultimate aim of beating a human world champion team by the year 2050. RoboCup soccer simulation is one of the categories in RoboCup soccer where the aspect of artificial intelligence is the main focus. This talk presents some winning approaches to the RoboCup soccer simulation competition using artificial intelligence technique. First, individual strategies are explained. The concept of chain-action is introduced in order to model the decision making process of an agent. Second, team-level strategy analysis is shown where a team strategy is represented using kick trajectories.
Tomoharu Nakashima (M'95) received the B.S., M.S., and Ph.D. degrees in engineering from Osaka Prefecture University, Osaka, Japan, in 1995, 1997, and 2000, respectively. He joined the Department of Industrial Engineering, Osaka Prefecture University, as a research associate in 2000, became an assistant professor in 2001. He was promoted to associate professor in the Department of Computer Science and Intelligent Systems in April 2005, and he was appointed as a professor in College of Sustainable System Sciences in April 2013. His current research interests include fuzzy-rule-based systems, RoboCup soccer simulation, machine learning application in industries, and healthcare information systems. He has served as a trustee member of SOFT in 2017-2019. He is now a secretary of IFSA council. For RoboCup soccer simulation, he has won three championships in the world RoboCup competitions. He is currently an executive committee member of Soccer Simulation League in RoboCup Federation. He also serves as a trustee member of RoboCup Japanese National Committee.
Recently, the declining labor force due to the low birthrate and longevity have been concerns. Since the shortage of labor is also a critical issue in infrastructure development and maintenance, the artificial intelligence (AI) techniques have been introduced to solve the issue. In this paper, we focus on the inspection and checking in constructing structures which have been requiring experiences and skills of experts. First, we formulate the tasks as a recognition problem in which the degree of reinforcement of the structure is classified by exploiting the potential relationships in the data observed in construction sites. In addition, we show a way to resolve inconsistencies in the data. Next, we propose two kinds of multi-class support vector machines (SVMs) which can exploit the relationships in the data. Finally, we evaluate the validity of the formulation and the classification results obtained by applying the proposed SVMs to some real datasets for one kind of large-scale structure.
Keiji Tatsumi received the bachelor's degree in engineering from Kyoto University in 1993, and the master's degree in information engineering from Nara Institute of Science and Technology in 1995, and Ph. D. degree in informatics from Kyoto University in 2006. He joined the Department of Engineering, Osaka University, as a research associate in 1998, and was promoted to an associate professor in 2009. His current research interests include metaheuristic method based on chaos dynamics, swarm intelligence, global optimization, and support vector machine.
In this talk, we will show the decision modeling we have been working on national level planning with the Electricity Generating Authority of Thailand (EGAT). Unit Commitment (UC) is the planning process for an electricity power system to provide the supply of electricity from the power generator to meet the forecasted demand. The decisions of the Unit commitment include which units to turn on/off, the energy generated from operating units, and the selection of the fuels. The uncertainty facing in the decision-making process is majorly from the errors of the forecasted demand. There is a fast introduction of renewable energy, especially from the Solar energy. This could bring down the demand and suddenly increase the demand if there is some cloud or rain. Moreover, the Electricity Vehicle (EV) will become major electricity consumption soon. The uncertainty of the charging time and energy could make the electricity system unstable. In addition, the error from forecasted fuel price could affect the decision of the fuel selection or generator unit selection. The integration of these uncertainties into the unit commitment becomes very important. We will discuss the modeling techniques that are suitable for solving the unit commitment with the uncertainty.
Chawalit Jeenanunta holds a B.Sc. in Mathematics and Computer Science, and M.Sc. in Management Science from University of Maryland. He received his Ph.D. in Industrial and Systems Engineering from Virginia Polytechnic Institute and State University. He joined Sirindhorn International Institute of Technology, Thammasat University, Thailand as a lecturer and now he is an associate professor. He was a chair of Management Technology curriculum, Head of School of Management Technology, Deputy Director for Building, Ground and Properties. He is also a head of Logistics and Supply Chain Systems Engineering Research Unit (LogEn) and the head of the Center for Technology Transfer of Industry 4.0. His Research interests are in area of applications of operations research, and innovation. He received research funding from Thailand Research Fund, Institute of Developing Economies and Japan External Trade Organization (IDE-JETRO), Economic Research Institute for ASEAN and East Asia (ERIA), PTT Logistics and Electricity Generating Authority of Thailand (EGAT).