基于时间序列的软件可靠性建模技术研究
本文首先介绍软件可靠性及其建模的一些基本概念，随后列举几种基于时间序列的软件可靠性增长模型，并分析他们的优缺点。最后着重介绍基于EM算法的软件可靠性预测模型，并用Matlab实现建模过程，通过所做的实验研究分析说明在软件可靠性建模领域中EM算法的可行性以及它的局限性。
关键词 软件可靠性 建模 时间序列 EM算法
毕业论文设计说明书外文摘要
Title Research of Software Reliability Modeling Based on Time Series
Abstract
With the coming of information era, the role that computer software plays in human activities is increasingly outstanding. The increasing dependence on software means that once the fault occurs, human being may have to confront with unacceptable loss, ranging from economic loss to even the loss of human life. Therefore, a new term ‘software reliability’ has raised public concern. Having been studied for several decades, software reliability becomes increasingly attention-grabbing instead of being forgotten. And it has already been a hot topic of discussion in modern society. As computer software has penetrated into every aspect of human activity, it is an urgent problem—how to improve software reliability in order to serve social production—for companies and researchers to address.
In this essay, firstly I will introduce the fundamental conception of software reliability and its modeling. After that, several kinds of software reliability growth models based on time series will be discussed and their advantages as well as disadvantages will be emphasized later. At last, I will state the main part of my research—software reliability prediction model based on EM algorithm according to my experiments and the result data obtained from it, including the importance of EM algorithm and its limitations.
Keywords software reliability model time series EM algorithm
目 次
1 引言（或绪论） 1
2 软件可靠性建模 2
2.1 软件可靠性 2
2.2 软件可靠性建模 3
2.3 基于时间序列的软件可靠性建模 4
3 使用EM算法进行软件可靠性建模 6
3.1 随机动态模型 6
3.2 卡尔曼滤波 7
3.3 EM算法 9
4 Matlab建模及实验结果分析 13
4.1 实验环境 13
4.2 Matlab建模 13
4.2 实验结果分析 17
结论 36
致谢 37
参考文献38
1 引言（或绪论）
随着信息时代的来临，对于人类来说，计算机软件的重要性与日俱增。对软件的依赖增大意着，一旦软件发生故障，可能会给人类生产活动带来难以接受的损失。
人类对飘渺太空、漫漫宇宙的探索从未停止，随着嫦娥三号在月球的着陆，中国人的奔月梦终于成为现实。这举世瞩目的成就就是得益于软件可靠性在航天航空领域的应用。然而，在我们看到这些荣耀的同时，软件故障带来的危害也不容忽视。比如，美国航天局耗资3.2亿美元打造的“火星气象卫星”Viking在接近火星表面时突然失踪，究其原因就是没有统一弹道计算的量纲。又如，2003年5月4日，俄罗斯宇宙飞船TMA1号，从宇宙空间站返回地球时，由于软件失效导致导航系统崩溃，使得在整个返回过程中与控制中心失联达11分钟，并且飞船的着陆地点与原定理论位置偏离460公里。再比如，1996年6月4日，历经10年才完工的ARIANE5火箭，竟然在首次发射升空40s后，由于攻角偏差，引起强烈的气动载荷，无可奈何下只能启动自毁装置引爆，造成的直接经济损失达到5亿美金。
上一篇：MOOC课程规划与设计研究