摘要
Je linsk i-M oranda(J-M)模型是基于失效间隔时间的较理论化的可靠性增长模型,用于软件系统测试阶段的可靠性度量。J-M模型存在的问题是假设条件理想化,实际应用环境难以满足,而且模型的失效间隔时间数据收集很困难,实用性差。该文提出一种改进的J-M可靠性增长模型,在J-M原模型基础上改进并引入更加合理的假设条件,改进后的模型基于缺陷计数,数据收集更容易。结合盐湖城奥运会信息系统软件测试数据对改进的模型进行检验,证明改进的模型预测有效性好,同时还保持了原模型的简单性。
The Jelinski-Moranda （J-M） model is a theoretical software reliability growth model based on the time between failures. The model is used to measure the reliability during system testing. However, the J-M model has ideal assumptions that are hard to satisfy and the times between failures are difficult to collect. This paper presents a modified J-M model with practical assumptions which are easy to meet, The modified model, based on fault counting, facilitates convenient data collection. The model was evaluated using integration testing data from the Information Systems for the Salt Lake Olympic Games. The results show that the modified J-M model can accurately predicate software reliability with a model that is as simple as the original model.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2006年第10期1759-1761,1766,共4页
Journal of Tsinghua University(Science and Technology)
基金
国家"十五"科技攻关项目(2002BA904B13)
