The game dynamics of social interaction: Algorithms and applications

该项目旨在开发数学模型来描述大量相互作用的战略代理人的动态。 由于在线社交网络取得了前所未有的成功，有关这些现象的详细数据越来越多。 博弈论模型提供了一个灵活的数学框架，纳什均衡应该描述这种模型的长期行为。 这项研究解决了几个基本问题，这种方法：选择多个纳什均衡？ 一个系统收敛到这样一个平衡点的速度有多快？ 如何影响这些动态？ 利用理论计算机科学的最新进展，将开发出有效的算法来回答这些问题。“社交网络”一词是指一群个体及其一对一的社会关系（可以是个人或经济性质的）。 很明显，这种网络的结构深刻地影响着整个社会的行为（特别是从经济的角度来看）。 由于缺乏有关社交网络结构和动态的详细数据，这种直觉从未变得实用或量化。 互联网，特别是在线社交网络的成功正在极大地改变这种情况。 该项目旨在开发迫切需要利用这种数据爆炸的数学模型和算法。 它将为经济分析工具在这一新的竞技场的应用开辟道路，从而促进利用在线网络的新途径。

This project aims at developing mathematical models to describe the dynamics of large populations of interacting strategic agents. Detailed data concerning such phenomena are becoming increasingly available thanks to the unprecedented success of online social networks. Game-theoretic models provide a flexible mathematical framework and Nash equilibria should describe the long term behavior of such models. This research addresses several fundamental problems with this approach: Which of multiple Nash equilibria is selected? How rapidly does a system converge to such an equilibrium? How can one influence these dynamics? Leveraging recent advances in theoretical computer science, efficient algorithms will be developed to answer such questions.The expression "social network" refers to a population of individuals together with their one-to-one social relationships (which can be of personal or economic nature). It is clear that the structure of such a network deeply affects the behavior of society as a whole (in particular, from the economic point of view). This intuition has never become practical or quantitative due to the lack of detailed data on the structure and dynamics of social networks. The Internet, and in particular the success of online social networking is dramatically changing this situation. This project aims at developing mathematical models and algorithms urgently needed to harness this data explosion. It will open the way to the application of economic analysis tools to this new arena, and thus facilitate new ways of exploiting online networks.