OPS5风格的专家系统工具

本文介绍了在Sun工作站和IBM—PC机上实现的一种新型的OPS5系统,重点描述其中的不精确推理和在产生式左部进行运算的功能实现。     

一种一致性的不确定性推理模型

将修正了的ＭＹＣＩＮ的可信度因子ＣＦ与ＰＲＯＳＰＥＣＴＯＲ的两种规则的不确定度量充分性和必要性度量ＬＳ,ＬＮ,放在ＢＡＹＥＳ框架下进行综合考虑,建立ＣＦ与ＬＳ,ＬＮ的函数关系,最后用实例说明了此模型的计算方法与主观ＢＡＹＥＳ方法的结果是一致的     

资本预算中的确定当量法之澄清

与风险调整折现率法不同,主观确定当量法无法得到一个投资项目的客观估值结果.即使给定决策主体,主观确定当量本身也难以同时满足可加性和可乘性,因而不能保证投资项目评估结果的逻辑一致性.对二叉树模型的无套利分析揭示了确定当量法、风险调整折现率法和风险中性概率法这三种估值方法之间的联系,并发现确定当量这一概念本质上应当是风险中性期望.在资本资产定价模型成立的条件下,给出了客观确定当量的计算公式.     

具有模糊系数的可能性线性规划

在LA Zadeh的可能性理论的基础上，讨论了具有梯形模糊系数的可能性规划问题，得到了一定置信水平下的经典等价条件，并用单纯形法求得其最优近似解．     

给排水工程专家系统（WTPES）不精确推理策略

从知识和证据的不确定性，可信度值的更新及传播等方面进行分析，介绍专家系统不精确推理技术在给排水工程中的应用情况，以ＷＴＰＥＳ软件包为例，介绍知识库结构及不精确推理模型的建立方法。     

目标类型识别的改进灰关联模型

针对防空作战目标类型识别的具体需求,分析了传统和广义灰关联模型在处理数据时存在的不足之处,建立了基于熵权法的改进灰关联目标类型识别模型。首先,为了提高空情数据的利用率,运用熵权法对数据熵值进行描述而后客观地赋予权重;其次,用置信度取代数据的绝对差值,更加准确地描述数据间相对差异的偏差情况;最后,对模型输出结果进行离散化处理,增强了模型的区分能力。用实例对模型进行检验,结果表明:该模型准确、简单且识别率高。     

多级教育发展规划决策支持系统——MESS

本文介绍了用于多级教育发展规划的决策支持系统MESS(Multi-LevelEducational-Programing Support System)。文中结合人工智能中的知识描述方法,将教育规划中各级教育部门的层次结构、各种辅助决策模型以及决策人的经验、偏好等表达为逻辑结构(事实和规则)的形式,并在此基础上进行决策问题求解路径的搜索及问题求题。     

Problems and promises: How to tell the story of a Genome Wide Association Study?

The promise of treatments for common complex diseases (CCDs) is understood as an important force driving large scale genetics research over the last few decades. This paper considers the phenomenon of the Genome Wide Association Study (GWAS) via one high profile example, the Wellcome Trust Case Control Consortium (WTCCC). The WTCCC despite not fulfilling promises of new health interventions is still understood as an important step towards tackling CCDs clinically. The ‘sociology of expectations’ has considered many examples of failure to fulfil promises and the subsequent negative consequences including disillusionment, disappointment and disinvestment. In order to explore why some domains remain resilient in the face of apparent failure, I employ the concept of the ‘problematic’ found in the work of Giles Deleuze. This alternative theoretical framework challenges the idea that the failure to reach promised goals results in largely negative outcomes for a given field. I will argue that collective scientific action is motivated not only by hopes for the future but also by the drive to create solutions to the actual setbacks and successes which scientists encounter in their day-to-day work. I draw on eighteen interviews.     

Simulation Methods for an Abductive System in Science

We argue that abduction does not work in isolation from other inference mechanisms and illustrate this through an inference scheme designed to evaluate multiple hypotheses. We use game theory to relate the abductive system to actions that produce new information. To enable evaluation of the implications of this approach we have implemented the procedures used to calculate the impact of new information in a computer model. Experiments with this model display a number of features of collective belief-revision leading to consensus-formation, such as the influence of bias and prejudice. The scheme of inferential calculations invokes a Peircian concept of ‘belief’ as the propensity to choose a particular course of action.

一种骨架型专家系统工具的结构与实现

本文论述了笔者所研制的一种骨架型专家系统工具的结构、功能和实观方法。