基于关系模式的知识表示方法

本文提出了一种专家数据库系统中知识库的设计方法。采用关系模式来描述知识库，用数据库方法存储管理知识库，以解决紧密耦合结构的专家数据库系统的设计问题，改善知识库管理的功能，克服专家系统不能有效地利用现有的使用常规数据库技术建造起来的数据库的局限。     

A CONNECTIONIST APPROACH TO KNOWLEDGE ACQUISITION

A peculiar kind of architecture for artificial intelligence systems, or AIA, is presented with the algorithm, an assembler, needed to allocate the physical resources among the well-formed formulas of the prepositional calculus that is its associated language. Such an architecture has been simulated on the 32 Transputers net of the D1S-CSCI, Università di Napoli. Together with AIA, the associated language, and the assembler, we present here the algorithms currently under development: namely a system that exploits AIA as a data-driven, nonaddressable architecture to store selectively in an associative memory only its significant interactions with an unpredictable environment. Because in AIA it is possible to store only the instructions, not the data, and because to do this we need to change the connections between the nodes, the very act of storing means altering selectively some of the previously stored program and thus altering permanently the actual system-environment interactions, i.e., the system learns something.     

模态逻辑推理的翻译方法

文中研究了模态逻辑推理的翻译法，即把模态逻辑公式按照一定的规则翻译成经典逻辑公式，再用传统的定理器进行推理，文中指出，该方法在理论上保持了正规命题模态逻辑的可判定性，还给出了一些试验结果，说明该方法实际可行的。     

A MIXED INTEGER DYNAMIC PROGRAMMING APPROACH TO A CLASS OF OPTIMAL CONTROL PROBLEMS IN HYBRID SYSTEMS

An optimal control problem for hybrid systems is formulated based on hybrid machine models. A practical approach to solving the problem, suitable for a class of hybrid systems, is presented. This approach consists of transforming the hybrid machine model into a dynamic programming model. Transition costs, in this latter model, are computed using mixed integer programs formulated based on the structure of the original hybrid machine. It is shown that the optimal solution for this dynamic program corresponds to the optimal control decision sequence in the hybrid system. Practical examples, inspired from process-oriented industry applications, are provided to illustrate the solution approach.     

一种基于PCA的图像表示与检索技术

本文提出了一种新的基于图像颜色分布特征和空间位置特征的图像表示方法以及相应的检索技术，首先使用一个颜色匀质谓词对图像进行多层分解，然后从分解得到的子图像中导出图像特征：“阈值子特征”和“颜色匀质标志子特征”，由于这种图像特征的给数将随着图像分解层数的增加而迅速膨胀，为此利用主分量分析法对其降维，图像的相似度量定义为归一化的子特征间的欧氏距离的线性组合，实验结果表明：使用本文提出的方法获得的图像检索结果能较好地同人们的期望结果保持一致。     

非单调性和不一致性的处理——一个基于线性逻辑的方法

本文提出一种基于线性逻辑的理论L1,它除了能较好地反映常识推理的非单调性外,还可以在矛盾存在的情况下继续组织有效的推理。     

A FUZZY REASONING APPROACH TO SUPPORT KNOWLEDGE ACQUISITION

Abstract

Acquisition of a user's computer based problem-solving skill is an important research area in human-computer interaction. This type of information usually can be derived by careful inspection on the actual dialog behavior. To implement such a knowledge reasoning agent, several crucial issues are pointed out and carefully inspected. These include: (a) appropirtate knowledge representation schema that are able to demonstrate the causal relationship between pairs of dialog events; (b) the formulation of a valid formula in calculating the overall knowledge index from the background information; (c) determination of the minimum sufficient number of dialog events required to form a discernible pattern; and (d) generalization of categories of performance patterns that can be applied to all types of application domains. A prototype reasoning agent based on the proposed methodology is constructed and its effectiveness is verified with the dialog events during UNIX operations.     

线性规划在知识获取中的应用

本文提出一个利用线性规划来实现从实例归纳学习的实用方法，简称为ＰＫＡ方法．该方法适用于许多分类专家系统的知识获取．     

从知识表示到表示:人工智能认识论上的进步

知识表示是对智能进行模拟的一个数学模型,然而它可以不是一个对智能本质的描述,特别是传统的符号主义知识表示离揭示人的智能行为发生的内在过程还有很大的差距,在神经科学和心理学的指导下,通过对智能行为的生理基础和心理过程的研究,遵循“解释智能”的思想,可以得到对知识的心智表示的新认识,这种表示观的不同,预示着人工智能方法论上的进步。     

一种进化模糊逻辑控制器的新方法

结合进化学习分类器的密歇根和匹兹堡方法的优点,首次将对单条控制规则的评价引入了模糊逻辑控制器（ＦＬＣ）的进化过程中,解决了匹兹堡类型的学习分类器系统“强化信息的带宽窄”的问题,实现了ＦＬＣ在控制器级和规则级的同时进化,控制器的控制规则数目也可以自由变化,实验结果表明新方法有较高的效率,优化的模糊控制器的结构简单,性能良好。     

知识的单形表示方法研究

本文基于形式逻辑的“三段论”以及代数拓扑学的三角剖分理论提出了一种新的知 识表示方法——单形表示法，并与其它一些知识表示方法进行了比较．它在表达能力、表达 方式及解释能力等方面都有明显的特点．     

一种基于十元格蕴涵代数的知识表示方法

不确定知识表示中常使用语言值来表达其真实程度和可信程度。为处理具有语言值可信度的不确定性知识表示问题,基于十元格蕴涵代数,建立一个十元语言值可信度因子知识表示模型,并实现了具有语言值可信度因子的知识推理。所提出的知识表示方法可以有效地表达具有可比性或不可比性的语言值不确定性信息。实例说明所提出的方法在表示知识时能够尽量模拟具有语言值信息的人类思维模式,有利于对不确定性知识的智能处理。     

THE USEFULNESS OF A MACHINE LEARNING APPROACH TO KNOWLEDGE ACQUISITION

This paper presents results of experiments showing how machine learning methods arc useful for rule induction in the process of knowledge acquisition for expert systems. Four machine learning methods were used: ID3, ID3 with dropping conditions, and two options of the system LERS (Learning from Examples based on Rough Sets): LEM1 and LEM2. Two knowledge acquisition options of LERS were used as well. All six methods were used for rule induction from six real-life data sets. The main objective was to lest how an expert system, supplied with these rule sets, performs without information on a few attributes. Thus an expert system attempts to classify examples with all missing values of some attributes. As a result of experiments, it is clear that all machine learning methods performed much worse than knowledge acquisition options of LERS. Thus, machine learning methods used for knowledge acquisition should be replaced by other methods of rule induction that will generate complete sets of rules. Knowledge acquisition options of LERS are examples of such appropriate ways of inducing rules for building knowledge bases.     

AN APPROACH TO REASONING ABOUT HYBRID SYSTEMS

A formal approach is presented for proving temporal properties of dynamic systems. Its main advantage is that it can be used to prove properties of hybrid systems, i.e. those whose state contains both discrete and continuous parameters. In contrast, most current temporal reasoning techniques are restricted either to purely discrete systems or to purely continuous systems. Our approach is based upon a new modeling technique called DMOD. A DMOD model of a system defines the causality relation between events in the system, using definite clauses, i.e. logic programs. Thereby, the problem of reasoning about hybrid systems is reduced to one of reasoning about the behavior of definite clauses. As these possess a simple proof theory, reasoning is substantially simplified.     

对象式逻辑程序设计

本文首先对逻辑程序设计与对象式程序设计进行一些比较，然后介绍对象式逻辑程序设计的基本原理、新进展、应用及目前存在的主要问题。     

状态逻辑型程序设计语言

本文提出了一种新的纯逻辑式子句型程序设计语言．文中给出了语言的语法，非形式语义，子句的过程解释和基于约束归结的推理系统．对该语言来说，程序包含三类变量：输入变量，输出变量和用于控制机器资源的程序变量；被程序定义的函数符号可用于构造项或子项，并且还可用作为谓词符号；不需要低效的最广合一．由于这些因素，一个子句集本身隐含了顺序，分支，迭代和递归多种控制结构使得容易构造高效的定理证明系统．这种语言将是一种有坚实理论基础的，高效的，实际有用的高级确定性语言．     

面向对象的三维壳体非线性有限元程序设计方法

面向对象的有限元程序设计可以大大提高程序的模块化和代码的重用性，本文将结合三维壳体单元模型的特点、介绍如何对其进行面向对象的程序设计，根据非线性有限元的求解步骤，把整个程序框架划分成一些基类，并派生出相应的子类，本文只对其中的几个类（包括它们的派生类）进行重点阐述，并说明它们之间是如何来传递消息，这包括单元类，材料类，非线性方程求解器类，总装类和分析类及它们的子类。     

用演化Agent方法处理整数线性规划问题

描述了一种演化ａｇｅｎｔ模型,并用这种模型来处理整数线性规划问题,该模型基于求整数一性规划问题的分枝定界方法,具有解并行,目标驱动,时间复杂性低等特点。