Exploring Avatamsaka Game and Other Behaviors under Socila Dilemmas

我们的研究项目总体上是为了探索Avatamsaka博弈和社会困境下的相关制度，以及开发一些网络上的模拟程序，特别是在与两难境遇相关的市场模型中。请注意，Avatamsaka游戏是由Y.Aruka在先前的JSPS项目编号10430004(1998-2000)中建议的。我们特别试图在进化博弈中检查突变及其对代理策略和适应度的影响。我们也洞察到了未知突变的潜在回报，但遗憾的是，这样的考虑还没有被充分分析。Avatamsaka博弈是这样一种博弈，它总是可以使任何策略在纳什均衡下都是最优策略。在我们的项目中，我们感兴趣的是参与者之间的各种形式的交互、信号交换、参与者的记忆和计算，以及当我们观察到具有社会困境的重复的两人博弈时，参与者在采用她的策略时的错误。如果我们利用基本的Dyn…在Nowak(1993)的基础上，我们可以很容易地通过计算机模拟来分析协同进化、突变和行动噪声对解的影响。我们考察了以下三种情况：(M=0)玩家永远不会记住过去的历史；(m=1)玩家只有对手的最后行为；(m=2)玩家有对手和她自己的最后行为。我们专注于一个特定的策略：(1)永远不要采取只有她自己承担损失的策略。(2)当玩家面临困境时，采取合作。(3)只要有合作，就采取合作。我们通常将这种战略称为PAVROV。在我们的背景下，PAVROV可能是由一个被认为是理性的玩家选择的，他会考虑说服她的对手改变她的叛逃，以实现相互合作。因此，这种类型的惩罚可以被解释为是在同情中进行的。我们称PAVROV为Avatamsaka游戏中的慈悲心内惩罚(Pwith C)。在m=2的情况下，正如秋山所展示的那样，事实证明最强的策略是PAVROV，即Pwith C。利他主义惩罚也在其他社会困境框架中得到承认，比如在Boyd等人(2003)中。在其他网络模拟的研究中，我们成功地实现了网络上信息引导下的非对称寡头博弈。这一结果发表在2003年7月9日至12日在意大利乌尔比诺举行的第15届IMGTA意大利博弈论与应用会议上。至于Avatamsaka游戏模拟，也在2003年8月8日至10日在美国马萨诸塞州波士顿波士顿大学举行的第13届国际心理学与生命科学混沌理论年会上发表。较少

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