The Graph Model for Conflict Resolution (GMCR): Reflections on Three Decades of Development
The Graph Model for Conflict Resolution (GMCR): Reflections on Three Decades of Development
复制标题
DOI：
10.1109/cscwd.2019.8791910
复制
发表时间：
2019-05
期刊：
影响因子：
--
通讯作者：
K. Hipel
中科院分区：
文献类型：
--
作者：
K. Hipel
通过追溯其三十多年的历史发展，突显了冲突解决图模型（GMCR）解决各种复杂的现实世界冲突情况的基本设计和固有能力，并强调了在人工智能（AI）时代和过度拥挤的世界中冲突加剧的未来有意义的扩展的巨大机遇。通过基于反映现实中实际发生情况的假设为 GMCR 构建良好的理论基础，讲述了一个引人入胜的故事，讲述了 GMCR 如何能够向大胆的新方向扩展，并利用早期元博弈分析和后来的冲突分析范式中构建的许多重要的遗留决策技术。例如，从其前身开始，GMCR 可以利用元博弈分析中提出的选项形式来有效记录冲突，以及在从冲突分析领域计算状态稳定性时定义类似象棋的行为的偏好引发技术和解决方案概念。论文中概述的关键思想是 GMCR 当前和预计能力的基础，包括开发四种不同的方法来处理传递性或不传递性偏好的情况下的偏好不确定性；用于描述冲突下多种人类行为的广泛解决方案概念；独特的联盟分析算法，用于确定特定决策者是否可以通过合作在争议中表现更好；追踪冲突随时间的演变；以及 GMCR 的矩阵公式，以提高计算稳定性时的计算效率，并在理论上将 GMCR 扩展到新的方向。讨论了决策支持系统的基本设计，该系统允许研究人员和从业者轻松地将 GMCR 的上述和其他进步应用于现实世界中棘手的争议。逆向工程被认为是 GMCR 的人工智能扩展，用于通过计算确定决策者所需的偏好，以达到理想的状态，例如所有国家大幅削减温室气体排放的气候变化协议。尽管 GMCR 已成功应用于解决许多不同领域中出现的争议，但美国和中国之间的一个简单的气候变化谈判冲突却被用来清楚地解释围绕 GMCR 的迷人历史旅程中提到的关键概念。
The fundamental design and inherent capabilities of the Graph Model for Conflict Resolution (GMCR) to address a rich range of complex real world conflict situations are put into perspective by tracing its historical development over a period spanning more than thirty years, and highlighting great opportunities for meaningful future expansions within an era of artificial intelligence (AI) and intensifying conflict in an over-crowded world. By constructing a sound theoretical foundation for GMCR based upon assumptions reflecting what actually occurs in reality, a fascinating story is narrated on how GMCR was able to expand in bold new directions as well as take advantages of many important legacy decision technologies built within the earlier Metagame Analysis and later Conflict Analysis paradigms. From its predecessors, for instance, GMCR could take advantage of option form put forward within Metagame Analysis for effectively recording a conflict as well as preference elicitation techniques and solution concepts for defining chess-like behavior when calculating stability of states from the realm of Conflict Analysis. The key ideas outlined in the paper underlying the current and projected capabilities of GMCR include the development of four different ways to handle preference uncertainty in the presence of either transitive or intransitive preferences; a wide range of solution concepts for describing many kinds of human behavior under conflict; unique coalition analysis algorithms for determining if a given decision maker can fare better in a dispute via cooperation; tracing the evolution of a conflict over time; and the matrix formulation of GMCR for computational efficiency when calculating stability and also theoretically expanding GMCR in new directions. The basic design of a Decision Support System for permitting researchers and practitioners to readily apply the foregoing and other advancements in GMCR to tough real world controversies is discussed. Inverse engineering is mentioned as an AI extension of GMCR for computationally determining the preferences required by decision makers in order to reach a desirable state, such as a climate change agreement in which all nations significantly cut back on their greenhouse gas emissions. Although GMCR has been successfully applied to challenging disputes arising in many different fields, a simple climate change negotiation conflict between the US and China is utilized to explain clearly key concepts mentioned throughout the fascinating historical journey surrounding GMCR.