作　　者： ; ; ; ; ;

机构地区： 国防科学技术大学信息系统与管理学院

出　　处： 《系统仿真学报》 2009年第14期4461-4466,共6页

摘　　要： 针对鲁棒决策中面临的想定空间探索困难的问题,结合探索性分析,提出了支持鲁棒决策的探索性试验设计;给出了想定空间生成框架,并依据该框架的构成,将想定生成和想定分析综合为想定探索框架,给出了一种想定生成和分析的形式化分析方法;同时,依据探索性分析的实验面临的问题,提出了层次化试验设计方法,给出了该方法的逻辑步骤,可以充分结合人的分析能力和计算机的计算能力进行交互式实验,以获取对问题的鲁棒解。 In order to solve the problem of exploring scenario space faced by robust decision-making, an exploratory analysis experimental design was introduced for supporting robust decision-making. First of all, a scenario space generation framework, PUCM framework for short was established, and its construction was analyzed. After that,＇ the scenario generation and scenario analysis were integrated into one framework, named scenario exploratory framework, which is a formalization analysis approach for scenario generation and analysis. Finally, a hierarchical experimental design technology was proposed, which could integrate the analysis ability of human and the computing capacity of computers, to deal with the issues in the experiments based on exploratory analysis, and introduce its logic process to guide the interactive experiments for identifying robust strategies.

关 键 词： 鲁棒策略 探索性分析 不确定性 想定 想定空间 试验设计

领　　域： [自动化与计算机技术] [自动化与计算机技术]