1 School of Systems Science, Beijing Normal University , Beijing 100875 , People's Republic of China.
2 F-power Inc. , Roppongi 1-8-7-2F, Minato, Tokyo 106-0032 , Japan.
Proc Biol Sci. 2019 Mar 27;286(1899):20190001. doi: 10.1098/rspb.2019.0001.
Sustaining cooperation among unrelated individuals is a fundamental challenge in biology and the social sciences. In human society, this problem can be solved by establishing incentive institutions that reward cooperators and punish free-riders. Most of the previous studies have focused on which incentives promote cooperation best. However, a higher cooperation level does not always imply higher group fitness, and only incentives that lead to higher fitness can survive in social evolution. In this paper, we compare the efficiencies of three types of institutional incentives, namely, reward, punishment, and a mixture of reward and punishment, by analysing the group fitness at the stable equilibria of evolutionary dynamics. We find that the optimal institutional incentive is sensitive to decision errors. When there is no error, a mixture of reward and punishment can lead to high levels of cooperation and fitness. However, for intermediate and large errors, reward performs best, and one should avoid punishment. The failure of punishment is caused by two reasons. First, punishment cannot maintain a high cooperation level. Second, punishing defectors almost always reduces the group fitness. Our findings highlight the role of reward in human cooperation. In an uncertain world, the institutional reward is not only effective but also efficient.
维持非相关个体之间的合作是生物学和社会科学的一个基本挑战。在人类社会中，这个问题可以通过建立奖励合作者和惩罚搭便车者的激励机制来解决。大多数先前的研究都集中在哪些激励措施能最好地促进合作。然而，更高的合作水平并不总是意味着更高的群体适应性，只有能导致更高适应性的激励措施才能在社会进化中生存。在本文中，我们通过分析进化动力学稳定平衡点处的群体适应性，比较了三种制度激励（奖励、惩罚和奖励与惩罚的混合）的效率。我们发现，最优制度激励对决策错误很敏感。当没有错误时，奖励与惩罚的混合可以导致高水平的合作和适应性。然而，对于中等和大的错误，奖励效果最好，应避免惩罚。惩罚的失败有两个原因。首先，惩罚不能维持高水平的合作。其次，惩罚破坏者几乎总是会降低群体适应性。我们的研究结果强调了奖励在人类合作中的作用。在不确定的世界中，制度奖励不仅有效，而且高效。
