灰色系统理论在软件可信性评估中的应用研究

【作者】 杨俊峰；

【导师】 李丹宁；

【作者基本信息】 贵州大学 ， 计算机应用技术， 2009， 硕士

【摘要】 近年来,随着可信计算(Trusted Computing)研究的不断深入,国内外对它的研究已经开始从硬件层面向软件等方向转变。可是面对越来越复杂的各种应用软件,软件危机、错误、缺陷、故障、失效、病毒、恶意代码入侵是屡见不鲜,而用传统的方法解决这些问题又不理想。因此,如何设计开发高可信软件以及进行软件可信性评估成了值得高度重视的问题,但国内外在这方面的研究甚少。由于不同类型的软件对软件可信性特性(特别是软件可信性属性)的要求是不同的,软件综合可信性是不确定的,对软件可信性的认识还处于灰色阶段。因此,本文以灰色系统理论(Grey System Theory)和可信计算(Trusted Computing)为基础、软件可信性知识为依托、我国重大研究计划“可信软件基础研究”为导向,运用灰色系统理论方法对软件可信性评估进行初步探索和研究。本文的主要内容概括如下:◆通过对软件可信性概念的分析研究,提出软件可信性模型,详细分析软件可信性特性:软件可信性属性、软件可信性的降低因素、软件可信性保障技术、软件可信性管理与配置、软件文档的可信性以及用户对软件的主观信任性等;讨论软件可信性评估指标体系建立的原则、方法及评估指标的标准化,并给出软件可信性评估的一般过程。◆在概述灰色系统理论知识的基础上,深入探讨灰色关联聚类法、GM(1,1)模型、灰色统计评估、灰色聚类综合评估等,并把这些方法理论成功地应用到软件可信性评估指标的选取、软件缺陷预估、软件文档可信性评估以及软件可信性综合评估中。◆运用灰色系统理论完成企业信息检索可信性综合评估指标体系、及其可信性综合评估模型的建立,并对该模型实例演示。结果表明,这些方法理论上先进、技术上可行。更多还原

【Abstract】 Recently, the research of trusted computing inside and outside our country has started to change from hardware level to software level with the research going deeper and deeper. But nowadays more and more problems are brought by various applied software such as limitation, errors, defects, faults, failures, virus and malicious codes, which might damage our systems, and these problems are not easily to be solved by the traditional ways. So how to design and develop trusted softwares and evaluate their creditability, are becoming important issues. However, there are few research about it inside and outside our country.Since different types of software have different requirements about characters of software creditability, especially the different property of software creditability, the overall creditability of the software is uncertain, the recognition of software creditability is in the grey area. So this paper presents an initial research on software creditability evaluation based on grey system theory, trusted computing and Chinese major research plan of trusted software.Firstly, this paper packs up the software creditability concepts inside and outside our country, especially presents the model frame of software creditability; as well as characters of software creditability: software creditability properties, the factors of software creditability to be reduced, software creditability safeguard technology, software creditability management and configuration, the creditability of software documents and so on; and analyses the actuality of software creditability evaluation inside and outside our country, and then puts forward principles and methods about setting up the evaluation index system of software creditability and makes it, and lastly presents a process of software creditability evaluation.Secondly, mainly researches grey relational clustering method, the model of GM(1,1), grey statistics evaluation and grey overall clustering evaluation. And these methods successfully are applied to select creditability evaluation indicators, forecast defects, evaluate documents’ creditability and evaluate overall creditability of software.Finally, sets up the overall evaluation index system and the evaluation model of the enterprise information retrieval creditability based on grey system theory, and show-how instance process of the model. The results indicate that the method is advanced in theory and practical in technology.更多还原