In our multi-agent model, agents generate wealth from repeated interactions for which a prisoner’s dilemma payoff matrix is assumed. Their gains are taxed by a government at a rate α. The resulting budget is spent to cover administrative costs and to pay a bonus to cooperative agents, which can be identified correctly only with a probability p. Agents decide at each time step to choose either cooperation or defection based on different information. In the local scenario, they compare their potential gains from both strategies. In the global scenario, they compare the gains of the cooperative and defective subpopulations. We derive analytical expressions for the critical bonus needed to make cooperation as attractive as defection. We show that for the local scenario the government can establish only a medium level of cooperation because the critical bonus increases with the level of cooperation. In the global scenario, instead full cooperation can be achieved once the cold-start problem is solved because the critical bonus decreases with the level of cooperation. This allows to lower the tax rate, while maintaining high cooperation.

中文翻译：

---

政府应该奖励合作吗？来自基于代理的财富再分配模式的见解

在我们的多智能体模型中，智能体从重复的交互中产生财富，其中假定了囚徒困境的收益矩阵。他们的收益由政府按税率征税α. 由此产生的预算用于支付管理成本并向合作代理人支付奖金，只有通过概率才能正确识别p. 代理在每个时间步根据不同的信息决定选择合作还是背叛。在当地情况下，他们比较了两种策略的潜在收益。在全球情景中，他们比较了合作和有缺陷的亚群的收益。我们推导出使合作与背叛一样具有吸引力所需的关键奖励的分析表达式。我们表明，对于当地情景，政府只能建立中等水平的合作，因为关键红利会随着合作水平的提高而增加。在全球场景中，一旦解决了冷启动问题，就可以实现全面合作，因为关键奖金会随着合作水平的提高而降低。这允许降低税率，同时保持高度合作。