|授業科目名||Artificial Intelligent Market Experiment|
|履修条件||Requirements to join in this lecture are not specified.|
|科目の目的・到達目標||We use a platform called "U-Mart system" in order to study Artificial Intelligent Market. This AI market simulator in essence is actually the same as the existing index futures trading at Osaka Stock Exchange (OSE). I t is noted that the idea of U-Mart preceded the actual one. This simulator at is origin equipped with a hybrid transaction of human and algorithmic agents. The latter can imitate a so-called high frequency transaction (HFT) by the acceleration experiment. So U-Mart is a progressive simulator to imitate the reality.
First of all, we should state our U-Mar project. Originally in 1998 U-Mart Project started as V-Mart(Virtual Mart). Now it however becomes re-nominated U-Mart. The U-Mart Project has published an English textbook (Shiozawa et. al. 2008) as one of Springer Series on Agent Based Social Systems in Spring 2008. The development of the U-Mart system during these 10 years rather was mainly engineers-driven. Now the U-Mart system is internationally recognized as a good platform for AI markets. Our project has now marked Version 4 which could be compatible with spot and futures market as well as batch and double auction.
Our U-Mart Lectures consists of two parts: one is this class. The other is the class for experiment, or training program. Both lectures normally are integrated into a single Aims and Scope of U-Mart lectures. So you are recommended to join into both classes of Artificial Intelligent Economics and Artificial Intelligent Market Experiment.
|授業の概要||The U-Mart System is an artificial intelligent market system to implement a virtual futures market with reference to the actual stock price index arbitrarily chosen, by the use of agent-based simulation techniques. In our experiments, we will have not only the training program for dealings as a human agent in the futures market but also for making a machine agent by JAVA programming. Our program is quit rudimentary only for beginners. . So we encourage you to join in our classes.|
|授業計画||The U-Mart learning will be constructed by the next parts.
I. The introductory lessons for training market trades as a human agent
II. The network experiments for human trading
III. The introductory lessons for JAVA programming for making a machine agent
IV. The network experiments for machine agents
The latter parts of III and IV seem advanced. The learning of the latter parts depend on the ability of participants. So we will mainly concentrate the first parts I and II in our lessons.
1. Introductory Guidance of the U-Mart System
2. (Prof. Yoshihiro Nakajima, Osaka City University) Play the elementary experiment
3. (Prof. Yusuke Koyama, Shibaura Institute of Technology) Play the traditional version of Itayose (Batch Auction)
4. (Prof. Kazuhisa Taniguchi, Kinki University) Learn the Analysis of the Market Experiment in the case of Itayose (Batch Auction)
5. (Prof. Naoki Mori, Osaka Prefectural University) Play the traditional version of Zaraba (Double Auction)
6. (Prof. Yoshihiro Nakajima) Learn the Analysis of the Market Experiment in the case of Zaraba (Double Auction)
7. A Short Tour to Tokyo Stock Exchange
8. A Short Tour to AI Laboratory of Tokyo Institute of Technology
10. (Prof. Yoshihiro Nakajima ) An elementary method to design the market simulation by the use of the Excel sheet
11. (Prof. Hiroshi Deguchi,Tokyo Institute of Technology) A New Application of Gaming Simulation (continued from the last lecture unit)
12. (Prof. Takashi Yamada, Yamaguchi University) The Analysis of Price Time Series (continued from the last lecture unit)
13. (Prof. Naoki Mori, Osaka Prefectural University) Try to make a new machine agent
14. (Prof. Takao Terano, Tokyo Institute of Technology) A New Application of Agent Base Simulation (continued from the last lecture unit)
15. Summing up
|評価方法||Two ways for grade the credit:
1. Do a paper on the given subjects.
2. Days attended
|テキスト・参考文献等||Y. Shiozawa, Y. Nakajima, H. Matsui, Y. Koyama, K. Taniguchi, and F. Hashimoto, Artificial Market Experiments with the U-Mart System, Springer ABSS Series Vol. 4, 2008, Springer-Verlag;
Kita, Hajime, Taniguchi, Kazuhisa, Nakajima, Yoshihiro (Eds.) Realistic Simulation of Financial Markets: Analyzing Market Behaviors by the Third Mode of Science, Springer Evolutionary Economics and Social Complexity Science, Vol. 4, forthcoming, 2017
|授業外の学習活動||Each virtual transaction experiment you should analyse the market result containing your own result. The analysis will improve your understand of the market transaction. The analytical method
will be taught during a series of lectures step by step.