权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

Studies on Qualitative Analysis of Dynamical Systems and its Application to Engineering Problem Solving

动力系统定性分析及其在工程问题解决中的应用研究

基本信息

批准号：
63550273
负责人：
NISHIDA Toyoaki
金额：
$ 1.15万
依托单位：
KYOTO UNIVERSITY
依托单位国家：
日本
项目类别：
Grant-in-Aid for General Scientific Research (C)
财政年份：
1988
资助国家：
日本
起止时间：
1988 至 1989
项目状态：
已结题

项目摘要

Qualitative reasoning plays an important role in human problem solving process. Given a complex problem, human experts first try to grasp qualitative nature of the problem by rough analysis. This kind of mental process has received much less attention than quantitative computation.The purpose of this research is to model and implement qualitative reasoning by human experts and to establish a well founded ground for engineering problem solving.The result of this research can be summarized as follows: 1. Studies on qualitative representation and reasoning about commonsense Knowledge about space. We have developed a method of reasoning about three-dimensional spatial configuration using a technique called potential model. We developed a potential minimization algorithm applicable to high dimensional parameter space concerning space such as position, size, shape, orientation, or neighborhood. 2. Studies on maintenance and utilization of incomplete information about quantity. We have developed a system which can answer questions by maximally making use of available information. We employed various kind of heuristic knowledge for this system. 3. Automated modeling by heuristic knowledge. We have developed a program which can automatically produce a model appropriate for a given problem by applying various model modification operators to basic model components. We make use of heuristic knowledge concerning causal analysis of underlying differential equations.

定性推理在人类问题解决过程中起着重要的作用。对于一个复杂的问题，人类专家首先试图通过粗略的分析来把握问题的定性。本研究的目的是建立和实现人类专家的定性推理模型，为工程问题的解决奠定良好的基础。本研究的成果概括如下：1.空间常识的定性表示和推理研究。我们开发了一种利用势模型技术对三维空间构型进行推理的方法。提出了一种适用于涉及位置、大小、形状、方向或邻域等空间的高维参数空间的势最小化算法。2.不完全数量信息的维护与利用研究。我们开发了一个系统，可以通过最大限度地利用现有信息来回答问题。我们在这个系统中使用了各种启发式知识。3.利用启发式知识进行自动建模。我们已经开发了一个程序，它可以通过对基本模型组件应用各种模型修改算子来自动生成适合给定问题的模型。我们利用启发式知识对基本的微分方程进行因果分析。