DMUU: Improving Decisions in a Complex and Changing World

DMUU：在复杂多变的世界中改进决策

• 批准号: 0345925
• 负责人: Robert Lempert
• 金额: $ 240万
• 依托单位: Rand Corporation
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2004
• 资助国家: 美国
• 起止时间: 2004-09-15 至 2012-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0345925&HistoricalAwards=false
• 关键词: DMUU; Improving; Decisions; Complex; Changing

This project will conduct fundamental research on two key questions important to the design and use of decision tools for supporting climate-change decisionmaking: (1) What are the best ways to represent uncertainty for decisionmakers and (2) What tools and methods work best in practice in providing these representations to decisionmakers? In particular, the project will strengthen the scientific foundations of robust decisionmaking (RDM), a promising new approach to computer-assisted support for decisions under conditions of deep uncertainty; that is, when decisionmakers are unsure of the system model or the prior probability distributions across the inputs to the system model(s). RDM helps decisionmakers to identify robust strategies whose satisfactory performance is largely independent of the resolution of most "unknowns" and in characterizing the residual deep uncertainties via their impact on the choice among strategies. RDM has the potential to significantly improve decisionmaking for climate-related and other policy areas. Interest in RDM is motivated by increased understanding of the processes of decisionmaking and the new capabilities of modern information technology which enable unprecedented interactions between groups of decisionmakers and tools for quantitative decision support. This research will draw on interactions with decisionmakers in two policy areas: (1) long-term planning for the management of water supplies by the California Department of Water Resources and (2) the design of scientific observation systems that could provide actionable warning of abrupt climate change. In each policy area, the project will employ integrated assessment models to characterize policy-relevant uncertainty in different ways (e.g., best estimates, probability densities, and vulnerabilities of robust strategies), perform Judgment and Decision-Making (JDM) experiments to determine how decisionmakers respond to different characterizations of uncertainty, deploy with decisionmakers tools that employ promising characterizations of uncertainty, and assess how these uncertainty characterizations perform in actual use. The project also will conduct basic research that identifies and evaluates the best statistical and other algorithms for supporting robust decisionmaking under a variety of conditions.This project will significantly enhance the scientific foundations of new approaches to climate-related decisions under conditions of deep uncertainty, and it will enrich understanding of decisionmaking processes under these ubiquitous conditions. The research will provide important new insights into the types of algorithms that can best be used to support these new decision approaches. The project will improve understanding of two important climate-related policy areas, water policy and abrupt climate change. Not only will the project's findings directly aid decisionmakers in these two domains, the project will improve methods to support decisionmaking under conditions of deep uncertainty useful in addressing a very wide range of public and private sector decision challenges. This award was supported as part of the Fiscal Year 2003 Human and Social Dynamics priority area special competition on Decision Making Under Uncertainty (DMUU).

该项目将对设计和使用决策工具支持气候变化决策至关重要的两个关键问题进行基础性研究：(1)为决策者表述不确定性的最佳方式是什么；(2)在向决策者提供这些表述方面，哪些工具和方法在实践中最有效？特别是，该项目将加强稳健决策的科学基础，这是在高度不确定的条件下，即当决策者不确定系统模型或系统模型投入的先验概率分布时，一种很有前途的计算机辅助决策的新方法(S)。RDM帮助决策者确定其令人满意的表现在很大程度上独立于大多数“未知数”的解决的稳健战略，并通过它们对战略选择的影响来表征剩余的深层次不确定性。区域发展管理有可能极大地改善与气候有关和其他政策领域的决策。增加对决策过程的了解和现代信息技术的新能力，使决策者群体和量化决策支持工具之间能够进行前所未有的互动，从而激发了人们对区域发展管理的兴趣。这项研究将借鉴与决策者在两个政策领域的互动：(1)加州水资源部对水资源供应管理的长期规划；(2)设计科学观测系统，为气候突然变化提供可行的警告。在每个政策领域，该项目将采用综合评估模型，以不同方式表征与政策有关的不确定性(例如，最佳估计、概率密度和稳健战略的脆弱性)，进行判断和决策(JDM)实验，以确定决策者如何对不确定性的不同表征作出反应，与采用前景看好的不确定性表征的决策者工具一起部署，并评估这些不确定性表征在实际使用中的表现。该项目还将进行基础研究，以确定和评估在各种条件下支持稳健决策的最佳统计和其他算法。该项目将大大加强在高度不确定条件下对与气候有关的决策采取新办法的科学基础，并将丰富对这些无处不在的条件下的决策过程的理解。这项研究将为最适合支持这些新决策方法的算法类型提供重要的新见解。该项目将提高对两个与气候有关的重要政策领域--水政策和气候突变--的了解。该项目的调查结果不仅将直接帮助这两个领域的决策者，该项目还将改进在高度不确定条件下支持决策的方法，有助于解决公共和私营部门非常广泛的决策挑战。该奖项作为2003财政年度人类和社会动力优先领域不确定性决策特别竞赛的一部分得到了支持。