Nagoya, Japan Dec 13-16, 2009

PRIMA 2009 Paciffic Rim Internarional Workshop on Multi-Agents

  • catch
  • aboutPrima
  • aboutPrima

Agent School


Date & Location

December 12th, 2009

Room Y (2nd Floor), Building #2 (campus map),
Nagoya Institute of Technology, Nagoya, Japan



  • 09:00 - 09:50


  • 09:50 - 10:00


  • 10:00 - 11:00

    Introduction to Artificial Market Reearch
    Kiyoshi IZUMI
    National Institute of Advanced Industrial Science and Technology (AIST), JAPAN

  • 11:15 - 12:15

    Multiagent Reinforcement Learning
    Sherief ABDALLAH
    Informatics Institute at the British University in Dubai, UAE

  • 12:15 - 13:30

    Lunch Time

  • 13:30 - 15:00

    Game Theory
    Makoto YOKOO
    Kyushu University, JAPAN

  • 15:00 - 15:30

    Coffee Break

  • 15:30 - 16:30

    Programming Rational Agents
    Delft University of Technology, NETHERLANDS

  • 16:45 - 17:45

    Advanced Software Design Techniques for Agent Systems
    Serge STINCKWICH Universite de Caen Basse Normandie, FRANCE

  • 17:45 -



How to apply/registraion

All participants must pre-register to attend the school. Please follow the registration information on PRIMA website and please let us know by sending an e-mail to us. The deadline for registration is Nov. 30th, 2009.

[Registration Fee]
Participants in PRIMA conference: FREE!
Participants in the Agent School ONLY: JPY 2,500


Abstract of Lectures

Introduction to Artificial Market Research
by Dr. Kiyoshi IZUMI

An artificial market is a virtual financial market on a computer. Agents, i.e., computer programs that play the role of virtual dealers, participate in an artificial market. Studies on artificial markets have attained some successes in market analysis in recent years. They have especially found new mechanisms for market phenomena such as financial bubbles that previous models could not adequately explain. They were also used to test existing economic theories such as efficient market hypotheses. This talk reviews a recent trend of artificial market researches towards market models tightly connected with actual financial markets. First, I will explain the electronized and computerized markets and the bottom-up approaches of market models as the the background of this research trend. Then I will introduce some recent artificial market researches; analysis of market bubbles, evaluation of automated trading, and integration with text mining.

Dr. Kiyoshi Izumi is the senior researcher of Digital Human Research Center, National Institute of Advanced Industrial Science and Technology. He earned his Ph.D. at University of Tokyo in 1998 and has worked in National Institute of Advanced Industrial Science and Technology since 1998. His current research topics include: 1) artificial market simulation; 2) human-flow simulation; and 3) a prediction market.

Multiagent Reinforcement Learning
by Dr. Sherief ABDALLAH

When an agent faces a new environment, where the outcome of the agent's actions are unknown, learning provides an excellent method for optimizing the agent's decision in such an environment. When there are more than one learning agents in the environment, then each agent needs to take into account other agents learning. In this talk I will review some of the recent work in multi-agent learning, comparing the pros and cons of each approach and the domains used for evaluation. I will also give a brief overview of the challenges faced by multi-agent learning research, with particular focus on multi-agent reinforcement learning.

Dr. Sherief Abdallah is Lecturer at the Institute of Informatics at the British University in Dubai, and an Honorary Fellow at the University of Edinburgh, UK. He holds MSc/PhD from the University of Massachusetts at Amherst, the United States. Dr. Abdallah's research focuses on developing reinforcement learning algorithms that are scalable and have some guarantee of convergence in a multi-agent context (particularly complex networks). He is also interested in applying machine learning and data mining to real and novel problems, including mobile devices, network management, and information retrieval.

Game Theory
by Dr. Makoto YOKOO

Game-theory and economics, in particular, mechanism design theory, can provide a solid theoretical foundation of autonomous agents and multiagent systems. Also,due to recent advances in Electronic Commerce, new application fields of autonomous agents and multiagent technologies, including Internet auctions, have been growing very rapidly. However, some basic concepts of game-theory and economics, such as equilibrium, rationality, uncertainty, etc., are not easily accessible to students with computer science or engineering backgrounds. This lecture gives an introductory tutorial of these basic concepts and their applications from the viewpoint of computer science.

Dr. Makoto Yokoo received the B.E. and M.E. degrees in electrical engineering, in 1984 and 1986, respectively, form the Univ. of Tokyo, Japan, and the Ph.D. degree in information and communication engineering in 1995, from the University of Tokyo, Japan. From 1986 to 2004, he was a research scientist of Nippon Telegraph and Telephone Corporation (NTT). He is currently a Professor of Information Science and Electrical Engineering, Kyushu University. His research interests include multi-agent systems, constraint satisfaction, and mechanism design among self-interested agents. He served as a general co-chair of AAMAS-2007, and as a program co-chair of AAMAS-2003. He received the ACM SIGART Autonomous Agents Research Award in 2004. He is on the board of directors of IFAAMAS.

Programming Rational Agents
by Dr. Koen HINDRIKS

Agent technology has been demonstrated to facilitate the design of systems that can operate in complex and dynamic scenarios. In particular, the concept of a rational agent has been proposed as a unifying concept to structure this design process. A rational agent is a complex computational process that combines such diverse capabilities as the processing of percepts and messages, maintaining and reasoning with a mental state, selecting actions and creating plans, and potentially also learning from experience. The engineering of such complex rational agent systems is a challenge, but we show that agent programming languages provide the tools to facilitate the design and development of such agents. The current state of the art and an overview of various agent programming languages is presented as a guide to the field. Some examples will be discussed to illustrate the use of agent programming to develop rational agents using the language GOAL.

Dr. Koen Hindriks is Assistant Professor at the Delft University of Technology. He studied computing science, and finished his PhD at Utrecht Univ. on agent programming languages. His research interests include common-sense reasoning, agent-oriented programming based on common sense concepts like beliefs and goals, and the verification and specification of agent programs. He has designed and developed several agent programming languages, including 3APL and GOAL. He is also interested in the design and development of negotiating agents, which involves among others research on representation, strategies and learning techniques that can be usefully applied in the development of such agents.

Advanced Software Design Techniques for Agent Systems

Agent technology has evolved rapidly over the last few years and is especially used when trying to solve problems in an uncertain world that changes frequently. This tutorial will deal with the general problem of adaptation and self-adaptation in the context of multiagent systems. Several methods from software engineering (meta-level and reflective architectures, aspect oriented programming) but also from cognitive sciences (anticipatory system, self-organisation and self-observation) will be presented and compared. Practical examples from diverse research fields like embedded systems, reconfigurable robotic systems or ambiant computing will also be provided. This lecture on the current state of the art of software engineering for agent systems will give an overview of this area without assuming any prior knowledge on the topic.

Dr. Serge Stinckwich is an associate professor of Computer Science at the Groupe de Recherche en Informatique, Image, Automatique et Instrumentation de Caen, Universite de Caen Basse Normandie (France). He is currently an inviting professor for two years in the MSI team (IFI) in Hanoi, Vietnam. His research interests include software architecture for multi-agent and robotic systems, Urban search and rescue robots, pedagogical robotics, anticipatory systems.



Jane Yung-jen HSU
National Taiwan University, Taiwan
[Personal web]

Hiromitsu HATTORI
Kyoto University, Japan
[Personal web]



Please e-mail to the following address:


Sponsors &

  • Arches, Inc.

  • Xacro Inc.

  • CSK Systems Chubu Corporation

  • Japan Agents Workshop and Symposium

  • Japanese Society for Artificial Intelligence

  • Information Processing Society of Japan

  • JSSST MACC (Japan Society for Software Science and Technology, SIG Multi-Agents and Cooperative Computation)

  • Nagoya Institute of Technology

  • Commemorative Organization for the Japan World Exposition '70

  • SCAT (Support Center for Advanced Telecommunications Technology Research)

  • Kayamori Foundation

  • Workshop on Multi-Agent and Cooperative Computation (MACC) in Japan Society for Software Science and Technology

  • Taiko Foundation