基于灰色神经网络的民航旅客信息系统可靠性模型研究

发布时间：2018-05-30 09:18

本文选题：民航旅客信息系统 + 软件可靠性预测　；参考：《中国民航大学》2017年硕士论文

【摘要】：民航旅客信息系统的7*24小时安全无故障运行,关系到航空公司的切身利益和航空旅客的出行体验,然而由于一系列的不可靠因素,国内外的民航旅客信息系统故障可谓是时有发生。因此,如何准确地评测系统的可靠性,维持系统的正常运行,是一个非常迫切的任务。本文尝试对大量的历史民航旅客信息系统运维监控数据进行分析,采用关键字提取以及正则表达式语义解析的方法,提取出数据中的可靠性失效数据,用于神经网络软件可靠性建模,准确预测失效数据,旨在减少可靠性预测中的潜在不可知因素,提高预测的精度。其主要工作如下:首先,分析和研究了可靠性数据提取方法,并设计了一种基于正则表达式的民航旅客信息系统可靠性数据提取方法,实现了程序化提取。其次,建立了民航旅客信息系统的BP神经网络和Elman神经网络两种可靠性模型,并利用建立的可靠性模型对民航旅客信息系统可靠性失效数据进行预测,验证了神经网络可靠性模型在民航旅客信息系统中的可行性,较好的解决了前期模型的系统适用性问题。再次,根据前两种模型的实验验证,发现了多用户量的系统软件中存在的可靠性数据的随机性和动态性比较强的问题,并针对此类情况提出一种将灰色模型和Elman神经网络相结合的灰色神经网络软件可靠性预测模型。最后,将三种模型的预测结果依据软件可靠性模型评价准则做了详细对比,结果表明基于灰色Elman神经网络的软件可靠性模型具有更好的预测精度和适用性,并在较大程度上提高了软件可靠性预测的准确性。
[Abstract]:The 7 / 24 hour safe and trouble-free operation of the civil aviation passenger information system is related to the vital interests of the airline and the travel experience of the airline passengers. However, due to a series of unreliable factors, At home and abroad, civil aviation passenger information system failures can be said to occur from time to time. Therefore, how to accurately evaluate the reliability of the system and maintain the normal operation of the system, is a very urgent task. This paper attempts to analyze a large number of operation and maintenance monitoring data of historical civil aviation passenger information system, and extracts the reliability failure data from the data by using the methods of keyword extraction and regular expression semantic analysis. In order to reduce the potential unknowable factors in reliability prediction and improve the accuracy of prediction, it is used for the reliability modeling of neural network software and accurate prediction of failure data. The main work is as follows: firstly, the reliability data extraction method is analyzed and studied, and a method based on regular expression is designed to extract the reliability data of civil aviation passenger information system. Secondly, two reliability models of civil aviation passenger information system, BP neural network and Elman neural network, are established, and the reliability failure data of civil aviation passenger information system are predicted by the established reliability model. The feasibility of the neural network reliability model in civil aviation passenger information system is verified, and the applicability of the previous model is well solved. Thirdly, according to the experimental verification of the first two models, the problem of randomness and dynamics of reliability data in multi-user system software is found. A software reliability prediction model based on grey neural network and Elman neural network is proposed. Finally, the prediction results of the three models are compared in detail according to the evaluation criteria of software reliability model. The results show that the software reliability model based on grey Elman neural network has better prediction accuracy and applicability. The accuracy of software reliability prediction is improved to a large extent.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V354;TP183

【参考文献】