Iyer Swami, Killingback Timothy
Computer Science Department, University of Massachusetts, Boston, Massachusetts, United States of America.
Mathematics Department, University of Massachusetts, Boston, Massachusetts, United States of America.
PLoS Comput Biol. 2016 Feb 29;12(2):e1004779. doi: 10.1371/journal.pcbi.1004779. eCollection 2016 Feb.
Cooperation in social dilemmas is essential for the functioning of systems at multiple levels of complexity, from the simplest biological organisms to the most sophisticated human societies. Cooperation, although widespread, is fundamentally challenging to explain evolutionarily, since natural selection typically favors selfish behavior which is not socially optimal. Here we study the evolution of cooperation in three exemplars of key social dilemmas, representing the prisoner's dilemma, hawk-dove and coordination classes of games, in structured populations defined by complex networks. Using individual-based simulations of the games on model and empirical networks, we give a detailed comparative study of the effects of the structural properties of a network, such as its average degree, variance in degree distribution, clustering coefficient, and assortativity coefficient, on the promotion of cooperative behavior in all three classes of games.
在从最简单的生物有机体到最复杂的人类社会等多个复杂层次的系统运行中，社会困境中的合作至关重要。合作虽然普遍存在，但从进化角度解释却极具挑战性，因为自然选择通常青睐非社会最优的自私行为。在此，我们研究了关键社会困境的三个范例中合作的进化情况，这些范例分别代表囚徒困境、鹰鸽博弈和协调博弈类别，研究对象是由复杂网络定义的结构化群体。通过在模型网络和实证网络上基于个体的博弈模拟，我们对网络结构属性（如平均度、度分布方差、聚类系数和 assortativity 系数）对这三类博弈中合作行为促进作用的影响进行了详细的比较研究。
