DRU: Supporting Group and Individual Decision Making

DRU：支持团体和个人决策

• 批准号: 0527650
• 负责人: Yoav Shoham
• 金额: $ 69万
• 依托单位: Stanford University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2005
• 资助国家: 美国
• 起止时间: 2005-09-15 至 2010-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0527650&HistoricalAwards=false
• 关键词: DRU; Supporting; Group; Individual; Decision

How should groups make decisions, and how can we support such decision-making processes? How should groups be formed to be effective? These questions motivate our research. To this effect we have formed a multi-disciplinary group of researchers who specialize in the areas of computer science, decision-theory, decision-support systems, game theory, and economics. In an attempt to make current tools and theoretical work more aligned with the needs of applications, we will focus on four important topics for group decision- making: (1) Obtaining and representing preferences of group members. This is a well-recognized bottleneck for decision-making because of the difficult cognitive requirements it places on individuals. (2) Efficiently solving games. Game theory is the primary mathematical tool used to model and understand group decision-making and group formation. Solutions to games are used to explain and to recommend social behavior, but computing such solutions is a notoriously difficulty problem. (3) Group and team formation. Given a group of people with different capabilities, how does one select the best team for handling a given task? Current methods are very limited in how they model such problems, e.g., they do not model negative interactions between group members. (4) Adaptive behavior. When the situation is dynamic and uncertain, decision makers need to adapt quickly - what algorithms should they use? The PIs have already made important contributions to these areas. This research program promises to push the envelopes of current scientific knowledge on the above issues, providing new and more efficient algorithms and more advanced models. Such results can have substantial benefit to society by helping government organizations and private corporations make better decisions, better adapt to change, and understand how to form more effective task forces and committees, and by providing them with tools that can support these decision making processes.

群体应该如何做出决策，我们如何支持这样的决策过程？应该如何组建团体才能有效？这些问题激发了我们的研究。为此，我们成立了一个多学科的研究小组，专门研究计算机科学、决策理论、决策支持系统、博弈论和经济学等领域。为了使当前的工具和理论工作更符合应用的需要，我们将重点研究群体决策的四个重要课题：（1）群体成员偏好的获取和表示。这是一个公认的决策瓶颈，因为它对个人提出了困难的认知要求。(2)有效地解决游戏。博弈论是用来模拟和理解群体决策和群体形成的主要数学工具。游戏的解决方案被用来解释和推荐社会行为，但计算这样的解决方案是一个众所周知的困难问题。(3)组和团队的形成。给定一组能力不同的人，如何选择最好的团队来处理给定的任务？目前的方法在如何对这些问题进行建模方面非常有限，例如，它们不模拟群体成员之间的消极互动。(4)适应性行为。当情况是动态的和不确定的，决策者需要快速适应-他们应该使用什么算法？PI已经在这些领域做出了重要贡献。该研究计划有望推动当前科学知识对上述问题的研究，提供新的，更有效的算法和更先进的模型。这些成果可以帮助政府组织和私营公司做出更好的决策，更好地适应变化，了解如何组建更有效的工作组和委员会，并为他们提供支持这些决策过程的工具，从而为社会带来实质性的好处。