标题： 博弈论分析复杂社会物理系统中的对抗性决策行为 显示英文标题

标题： Game-Theoretic Analysis of Adversarial Decision Making in a Complex Sociophysical System

摘要： 我们将博弈论应用于两个竞争代理团队在复杂网络中的数学表示，其中每一方操纵其资源并削弱对方的能力取决于他们内部同步决策的能力，同时超越对方。 这种对抗性社会物理系统的表示在商业、体育和军事等多种情境中都有应用。 具体而言，我们在这里结合了两种基于物理的模型，即Kuramoto模型用于表示决策周期，以及多物种Lotka-Volterra系统的改编用于资源竞争。 对于复杂网络，我们使用Barabási-Alberts无标度图的不同变体，改变资源在图中心和外围之间的初始分布。 我们将Nash主导博弈剪枝作为均衡解，以有效探索动态决策树。 在各种情景中，我们发现纳什解，其中最初将资源集中在外围的一方可以持续竞争以取得胜利，除非双方存在不对称性。 当结构优势有限时，获胜者如何保持对对方决策状态的领先变得至关重要。 显示英文摘要

摘要： We apply Game Theory to a mathematical representation of two competing teams of agents connected within a complex network, where the ability of each side to manoeuvre their resource and degrade that of the other depends on their ability to internally synchronise decision-making while out-pacing the other. Such a representation of an adversarial socio-physical system has application in a range of business, sporting, and military contexts. Specifically, we unite here two physics-based models, that of Kuramoto to represent decision-making cycles, and an adaptation of a multi-species Lotka-Volterra system for the resource competition. For complex networks we employ variations of the Barab\'asi-Alberts scale-free graph, varying how resources are initially distributed between graph hub and periphery. We adapt as equilibrium solution Nash Dominant Game Pruning as a means of efficiently exploring the dynamical decision tree. Across various scenarios we find Nash solutions where the side initially concentrating resources in the periphery can sustain competition to achieve victory except when asymmetries exist between the two. When structural advantage is limited we find that agility in how the victor stays ahead of decision-state of the other becomes critical.