Visual Hyperspace Pareto Frontier (VHPF) for Multiattribute Decision-Making

用于多属性决策的视觉超空间帕累托前沿 (VHPF)

• 批准号: 0600263
• 负责人: Kemper Lewis
• 金额: --
• 依托单位: SUNY at Buffalo
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2006
• 资助国家: 美国
• 起止时间: 2006-05-15 至 2009-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0600263&HistoricalAwards=false
• 关键词: Visual; Hyperspace; Pareto; Frontier; VHPF

In this project we advance a research plan to develop methods that will enable a decision-maker to make interactive, real-time trade-off investigations for multi-attribute design problems, under certainty or uncertainty. These trade-off investigations will focus on identifying which design candidates best satisfy the performance objectives (as represented by the multiple attributes), which design candidates best satisfy some subset of performance objectives, and which design candidates remain 'good' candidates, even under uncertainty related to incompleteness of knowledge or random factors in the original design problem. The Visual Hyperspace Pareto Frontier (VHPF) will be developed to enable real-time decision-making, to address the complexity involved in any large multiattribute problem, and to accommodate the incorporation of uncertainty considerations. The long-term goals of the project are 1) to support decision-making in design by providing a visualization-based tool to meaningful explore multidimensional performance space in multi-attribute design, 2) to provide a lossless dimension blending technique, and 3) to integrate the blending technique with multidimensional visualization. The broader impacts are expected from the capability how data can be synthesized and processed appropriately so as to derive use, particularly when coming from many different disciplines, such as for structure determination for genomics, or entertainment simulation, or engineering analysis of complex systems.

在这个项目中，我们提出了一个研究计划，开发方法，使决策者进行互动，实时权衡调查多属性设计问题，在确定性或不确定性。这些权衡调查将集中于确定哪些设计候选最好地满足性能目标（由多个属性表示），哪些设计候选最好地满足性能目标的一些子集，以及哪些设计候选仍然是“好”的候选，即使在与原始设计问题中的知识或随机因素的不确定性相关的情况下。将开发视觉超空间帕累托边界（VHPF），以实现实时决策，解决任何大型多属性问题所涉及的复杂性，并适应纳入不确定性因素。该项目的长期目标是：1）通过提供一个基于可视化的工具来支持设计中的决策，以有意义地探索多属性设计中的多维性能空间; 2）提供一种无损的维度混合技术; 3）将混合技术与多维可视化相结合。更广泛的影响预计来自如何适当地合成和处理数据以获得使用的能力，特别是来自许多不同学科的数据，例如基因组学的结构确定，或娱乐模拟，或复杂系统的工程分析。