| Co-Investigator(Kenkyū-buntansha) |
ODA Sobei Kyoto Sangyo University, Faculty of Economics, Professor, 経済学部, 教授 (40224240)
NAMATAME Akria National Defence Academy, School of Electrical and Computer Engineering, Professor, 情報工学科, 教授
AOKI Nobuo Chuo University, Faculty of Commerce, Professor, 商学部, 教授 (60087020)
KOYAMA Yuhsuke Interdisciplinary Graduate School of Science and Engineering, Computational Intelligence and Systems Science, Research Associate, 総合理工学研究科, 助手 (80345371)
AKIYAMA Eizo Tsukuba University, Institute of Policy & Planning Sciences, Lecturer, 社会工学系, 講師 (40317300)
|
|---|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2003: ¥1,200,000 (Direct Cost: ¥1,200,000)
Fiscal Year 2002: ¥2,400,000 (Direct Cost: ¥2,400,000)
|
|---|
| Research Abstract |
Our research project in general has been arranged for exploring Avatamsaka Game and the related systems under social dilemmas as well as developing some simulation programs on the net-work particularly in the market models associated with dilemma. Note that Avatamsaka game was suggested by Y.Aruka in the previous JSPS project No.10430004(1998-2000). We particularly tried to examine mutation and its effects on agent strategy and fitness in the evolutionary game. We also had an insight on potential pay-offs with unknown mutation, but such a consideration, to our, regret, has not been fully analyzed.
Avatamsaka game is the game which can always make any strategy an optimal strategy in terms of Nash equilibrium. In our project, we were interested in the various forms of interaction of players, signal exchanges, player's memory and calculation, and player's mistakes on adopting her strategy when we observed the iterated two-persons games with social dilemmas. If we utilized a fundamental dyn
… More
amic equation for evolutionary game in terms of Nowak(1993), we could easily analyze a set of the effects of co-evolution, mutation, and action-noise on the solutions by computer simulation. We examine the following three cases: (m=0) Players never have memory on the past history; (m=1) Player only has the opponent's last behavior; (m=2) Players has the last behaviors of both the opponent and herself. We focus on a particular strategy: (1)Never adopt such a strategy that only herself takes loss. (2)Adopt cooperation when player is faced to dilemmas. (3) Adopt cooperation as long as cooperation is held. We usually call such strategy PAVROV. In our context, PAVROV may be chosen by a supposed reasonable player who considers well to persuade her opponent to change her defection to realize mutual cooperation. Thus this type of punishment may be interpreted to be done within compassion. We call PAVROV a Punishment within Compassion(PwithC) in Avatamsaka game. In the case of m=2, as Akiyama showed, it turns out that the strongest strategy is PAVROV, namely, PwithC. The altruistic punishment is also acknowledged in other social dilemma frameworks like in Boyd et al.(2003). Thus our research can contribute the subject of evolution and altruism.
In the other research on net-work simulation, we were successful to achieve an Asymmetric Oligopoly Game with Information Guidance on the net work. This result was presented in XV IMGTA Italian Meeting on Game Theory and Applications Urbino, (Italy), July 9-12, 2003. As for Avatamsaka game simulation, also presented in the 13th Annual International Conference, The Society for Chaos Theory in Psychology & Life Sciences, Boston University, Boston, MA, USA, August 8-10, 2003. Less
|
