Liang Xiannuan, Xiao Yang
Department of Computer Science, The University of Alabama, Tuscaloosa, AL 35487-0290, USA.
IEEE Trans Syst Man Cybern B Cybern. 2010 Jun;40(3):683-93. doi: 10.1109/TSMCB.2009.2034976. Epub 2009 Nov 20.
In this paper, inspired by the society of animals, we study the coalition formation of robots for detecting intrusions using game theory. We consider coalition formation in a group of three robots that detect and capture intrusions in a closed curve loop. In our analytical model, individuals seek alliances if they think that their detect regions are too short to gain an intrusion capturing probability larger than their own. We assume that coalition seeking has an investment cost and that the formation of a coalition determines the outcomes of parities, with the detect length of a coalition simply being the sum of those of separate coalition members. We derive that, for any cost, always detecting alone is an evolutionarily stable strategy (ESS), and that, if the cost is below a threshold, always trying to form a coalition is an ESS (thus a three-way coalition arises).
在本文中，受动物群体行为的启发，我们运用博弈论研究用于检测入侵的机器人联盟形成问题。我们考虑在一个由三个机器人组成的群体中形成联盟，这些机器人在一条封闭曲线回路中检测并捕获入侵目标。在我们的分析模型中，如果个体认为其检测区域过短，以至于无法获得高于自身的入侵捕获概率，那么它们就会寻求联盟。我们假设寻求联盟存在投资成本，并且联盟的形成决定了胜负结果，联盟的检测长度仅仅是各个联盟成员检测长度之和。我们推导得出，对于任何成本而言，始终单独进行检测是一种进化稳定策略（ESS），并且，如果成本低于某个阈值，始终尝试形成联盟是一种ESS（从而会出现三方联盟）。
