Exploring Avatamsaka Game and Other Behaviors under Socila Dilemmas

我们的研究项目一般已安排探索化身游戏和社会困境下的相关系统，以及开发一些网络上的模拟程序，特别是在与困境相关的市场模型。请注意，Avatamentum游戏是由Y.Aruka在之前的JSPS项目No.10430004（1998-2000）中提出的。我们特别试图研究突变及其对进化博弈中代理策略和适应性的影响。我们还对未知突变的潜在收益有了深入的了解，但令人遗憾的是，这样的考虑还没有得到充分的分析。Avatamanti博弈是一种总能使任何策略成为纳什均衡的最优策略的博弈。在我们的项目中，我们感兴趣的是各种形式的互动的球员，信号交换，球员的记忆和计算，以及球员的错误采取她的策略时，我们观察了迭代二人游戏的社会困境。如果我们利用一个基本的动态 关于我们 利用Nowak（1993）的演化博弈方程，通过计算机模拟，可以很容易地分析协同演化、变异和行动噪声对解的影响。我们研究了以下三种情况：（m=0）参与者对过去的历史没有记忆;（m=1）参与者只有对手的最后一次行为;（m=2）参与者有对手和自己的最后一次行为。我们专注于一个特定的策略：（1）永远不要采取这样一个策略，只有她自己承担损失。（2）当玩家面临困境时采取合作的策略。(3)只要有合作，就采取合作。我们通常称这种策略为PAVRENT。在我们的背景下，PAVbind可能是由一个假定的合理的参与者选择的，他考虑说服她的对手改变她的背叛，以实现相互合作。因此，这种类型的惩罚可能被解释为在同情中进行。我们称之为Pavatam游戏中的PwithC惩罚。在m=2的情况下，正如Akiyama所展示的那样，最强的策略是PAV，即PwithC。利他主义惩罚在其他社会困境框架中也得到了承认，如Boyd等人。（2003年）的报告。在网络仿真的其他研究中，我们成功地实现了一个具有信息引导的非对称寡占博弈模型。这一结果发表在2003年7月9日至12日在乌尔比诺（意大利）举行的第十五届IMGTA意大利博弈论和应用会议上。至于阿凡达游戏模拟，也在第13届年度国际会议上提出，心理学和生命科学混沌理论学会，波士顿大学，波士顿，MA，美国，2003年8月8日至10日。少

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