航空机载软件缺陷知识库框架  

The Defect Knowledge Base Framework for Airborne Equipment Software

作　　者：张贺[1] 王世海[1] 刘斌[1] 杨顺昆[1] 余正伟[1] 秦蕾[1] 

机构地区：[1]北京航空航天大学可靠性与系统工程学院,北京100191

出　　处：《测控技术》2013年第1期99-103,共5页Measurement & Control Technology

基　　金：海外优秀人才培养基金

摘　　要：提高航空机载软件质量成为当前一个亟须解决的问题。建立软件缺陷知识库对于进行有效的软件质量评价及软件故障预测,识别易于出现缺陷的软件模块,提高软件测试效率和软件质量,都能起到重要作用。提出了一个基于机器学习和产生式系统推理相结合的航空机载软件缺陷知识库构建方法和相应的框架,该框架还包含软件缺陷度量元选取标准、选取清单,以及缺陷信息统计要求、分析方法。在此框架的基础上,利用实际测评工作中积累的大量航空机载软件缺陷数据,构建了一个统一、规范的软件缺陷知识库,并通过该知识库给出了缺陷预防信息,从而对航空机载软件全寿命周期进行了有效指导。Due to the importance of the airborne equipment software （AES）, a lot of attentions have been drawn into here. How to improve the quality on it, however, is still a challenge for all the researchers in this area. There is always a difficulty to access to a certain kind of knowledge from a huge amount of data, when experts make evaluations on AES＇s quality or predictions on AES＇s defects. From the AES testing projects, a large num ber of defect data is achieved. Building a defect knowledge base with unified, standardized and effective man agement AES is a definitely valuable work. This knowledge base is highly essential to make accurate evalua tions on the quality, predictions on the defects, identifications on the faultprone modules, and great reductions on the human cost in testing for AES. As the first step for building a knowledge base, first of all, the metrics and the analysis methods for the invalid data employed in AES knowledge base are designed and described. Af terwards, a framework on how to build an AES knowledge base is proposed, a combination mechanism is pro posed by involving machine learning technology and production system, in which, in order to provide the in structions for AES lifecycle development and maintenance.

关 键 词：航空机载软件 知识库 度量元 机器学习 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]