基于Bayes方法的软件运行剖面建模  

The New Algorithm of Evaluating SOP Based on Bayes Method

作　　者：李宗民[1] 刘玉杰[1] 白成刚[2] 

机构地区：[1]石油大学(华东)计算机与通信工程学院 [2]北京航空航天大学自动控制系,北京100083

出　　处：《计算机工程与应用》2004年第10期34-36,84,共4页Computer Engineering and Applications

基　　金：国家863高技术研究发展计划(编号:2001AA113192);航空基础科学基金(编号:01F51025)资助

摘　　要：软件运行剖面对于软件可靠性和软件测试具有重要意义,关于软件运行剖面建模,已提出多种方法。但是这些方法都未考虑先验信息,在现场测试数据匮乏的情况下,将这些先验信息丢弃不用,难以对实际运行剖面作出适当的估计。论文讨论了基于Bayes方法的软件运行剖面建模,该方法能够把先验信息与实际测试信息很好地结合起来,并给出了相应的算法。经仿真验证,基于Bayes方法的软件运行剖面建模方法能够较好地估计软件的实际运行剖面。SOP(software operational profile)is important for software reliability and software test.There have been many SOP models up to date.However,the prior information is not used in these models.When the testing data is not enough,it is difficult to evaluate pragmatic SOP properly.In this paper,the static model of SOP is improved.Bayes method is used to combine prior information with factual testing information.The paper gives a general algorithm of evaluating SOP.Through the simulation,it shows that the new method is better than traditional statistical method.

关 键 词：软件运行剖面 BAYES方法 Dirichlet分布 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]