中国科学院软件研究所机构知识库
Advanced  
ISCAS OpenIR  > 中科院软件所  > 中科院软件所
题名:
基于因果网络的诊断系统的研究
作者: 彭国强
答辩日期: 1998
专业: 计算机软件
授予单位: 中国科学院软件研究所
授予地点: 中国科学院软件研究所
学位: 博士
关键词: 诊断推理 ; 扩充因果网络 ; 诊断系统 ; 神经网络 ; 因果网络学习
摘要: 诊断问题求解是人工智能应用最早且直至今日仍是最活跃的应用领域之一。因果网络也是近年来人工智能研究的热点,它具有联接机制与符号机制两者的优点。因此,用因果网络能构造出功能更强的诊断系统。本文综述了诊断问题和因果网络的研究与发展,讨论了诊断推理的一些方法,并分析了这些方法的优缺点。评述了混合神经网络的研究工作。提出了扩充因果网络以及基于扩充因果网络的诊断推理模型和扩充因果网络的进化学习算法。最后,进一步给出了一个基于扩充因果网络的诊断系统模型。本论文就因果网络、基于因果网络的诊断推理以及因果网络的学习和基于因果网络的诊断系统模型几方面问题,进行了深入的研究。主要工作如下:1. 提出扩充因果网络。在noisy-or因果网络的基础之上,提出合作集与合作结点的概念(见第三章)。因果网络中加入这些合作结点和合作集,使因果网络得以扩充,即为“扩充因果网络”。实际当中,如果不考虑合作集和合人结点,就可能使诊断推理不精确或者不正确,导致错误的诊断结果。2. 给出基于扩充因果网络的诊断推理模型。首先用模糊推理方式建立诊断目标,直接把扩充因果网络作为神经网络结构,建立求结点激活值规则,通过神经计算达到平衡时的状态,给出诊断结果,并用反绎推理过程描述依断结果。然后,合出诊断求解步骤和所求解的个数的计算方法。最后是一些实验与分析以及对该模型的一些讨论,实验也证明了该模型的有效性和与其它诊断推理方法相比所表现出的优越性。3. 提出扩充因果网络的学习方法。在总结现有因果网络的学习方法的优缺点之后,给予出扩充因果网络的进化学习方法。讨论了该学习方法的缺点并给出相应的解决方法。该方法不仅能同时学习扩充因果网络的权值与结构,还能使扩充因果网络动态地适应变化的环境,实验也证明了该学习方法的有效性。4. 进一步给出一个基于扩充因果网络的诊断系统模型。该模型主要以电路故障诊断为背景,以第三、五和六章工作为基础建立的。设备的组件和其子组件构成设备的不同层次,诊断推理可以在不同层次上进行,每一层次对应因果网络的相应结点层。该模型所提供的诊断方法是以基于因果网络的分层方式进行的。对于给定诊断设备,用户可以根据需要选择诊断层次。此外,由于因果网络所具有的优势,使得该模型既具有较强的处理复杂问题的能力,又有较强的知识表示能力。实验与分析中,进一步讨论了该诊断系统的诊断方法和所具有的优势。
英文摘要: Diagnostic problem solving is one of the applied fields to be applied earliest and still be most active today in artificial intelligence applications. Causal networks, which have both the advantages of the connectionist and symbolic, are also the hot topics over the decade in artificial intelligence circle. So diagnostic systems constructed by using causal networks will be much stronger. This dissertation reviews the developments and researches of diagnosis and causal networks. It discusses some methods for diagnostic reasoning, analyses the advantages and disadvantages of these methods and describes some researches on hybrid neural networks. It gives Extended Causal Networks(ECN), a diagnostic reasoning model based on ECN and a learning method for ECN. Finally, it presents a model of diagnostic system based on ECN. This dissertation deals deeply with several important problems existed in causal networks, the diagnostic reasoning based on causal networks, the learning for causal networks and diagnostic systems based on causal networks. The following are the main researches. 1. Extended Causal Networks. In causal networks, the definitions of coaction sets and coaction nodes are given. When these coaction sets and coaction nodes are added as nodes to causal networks, the causal networks will be extended to be EXtended Causal Networks. Without thinking of these new nodes in practice, it may influence the accuracy of reasoning, and even give rise to incorrect reasoning to give wrong diagnostic results. 2. A diagnostic reasoning model based on ECN. We first use fuzzy reasoning method to construct the objective function, directly use ECN as neural networks to give the activation rules and give the diagnostic results according to the equilibrium of the neural computation and use abduction to describes the results, then give a method of computing the number of the solutions for the diagnostic reasoning. Finally, some experiments, analysis and discussion are given. The experiments show the effectiveness and advantages to compare with other diagnostic reasoning methods. 3. The learning method for ECN. After reviewing the advantages and disadvantages of the existing causal network learning methods, an evolutionary learning method for ECN are given. The disadvantages of the learning method for ECN are discussed and some solutions to these disadvantages are presented. This learning method is capable of not only simultaneously acquiring both the topology and weights of ECN but also dynamically adapting a ECN to a changing environment. The experiment results show the effectiveness of the proposed method. 4. A model of diagnostic system based on ECN. This model, based on the work of Chapter 3, Chapter 5 and Chapter 6 is constructed mainly for fault diagnosis of digital circuits. A device can be divided into different layers according to the components and the subcomponents. Each layer corresponds to the layer of the causal network. This model give the diagnostic method in the way based on causal networks with different layers to be selected. For a given device, the users can choose the layer they want. Because of the advantages of causal networks, the model not only have much stronger ability to process complex problems, but also have much stronger ability for knowlege representations. In the experiment and analysis, the diagnostic methods and advantages of the diagnostic system are furher discussed.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.iscas.ac.cn/handle/311060/5912
Appears in Collections:中科院软件所

Files in This Item:
File Name/ File Size Content Type Version Access License
N98839.pdf(2568KB)----限制开放-- 联系获取全文

Recommended Citation:
彭国强. 基于因果网络的诊断系统的研究[D]. 中国科学院软件研究所. 中国科学院软件研究所. 1998-01-01.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[彭国强]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[彭国强]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2017  中国科学院软件研究所 - Feedback
Powered by CSpace