资源描述:
《基于蚁群算法的软件可靠性模型参数估计方法》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库。
1、基于蚁群算法的软件可靠性模型参数估计方法摘要:由于软件可靠性模型大多是非线性模型,导致其参数难于估计。总结了常用的软件可靠性模型的参数估计方法,提出一种基于蚁群算法的可靠性模型参数估计方法。通过对musa软件可靠性模型分类方案中三个不同类型模型(g.o模型、weibull模型以及m.o模型)的实验,发现本算法对不同模型具有很好的适应性,解决了应用传统数值计算方法时的无法收敛问题;与粒子群算法相比,本算法的收敛速度比粒子群算法快一倍以上,且对于部分实验对象的拟合结果精度比粒子群算法高一个数量级以上。关键词:蚁群算法;软件可靠性
2、模型;参数估计estimatingparametersofsoftwarereliabilitymodelsbyantcolonyalgorithmzhengchang.you1*,liuxiao.ming1,huangsong21.instituteofcommandautomation,plauniversityofscienceandtechnology,nanjingjiangsu210007,china;2.engineeringinstituteofcorpsofengineers,plauni
3、versityofscienceandtechnology,nanjingjiangsu210007,chinaabstract:softwarereliabilitymodelingisoneofthebasementsofsoftwarereliabilityengineering.mostsoftwarereliabilitymodels’parametersarehardtoestimate,astheyarenonlinearfunctions.themostwidelyusedmethodsforparamete
4、rsestimatingofsoftwarereliabilitymodelshavebeensummarized,andanewapproachbasedonantcolonyalgorithmisproposedinthispaper.experimentswiththreetypicalmodels-g.omodel,weibullmodelandm.omodel-showthatthisalgorithmdemonstratesgoodapplicability.andtheresultsdemonstratetha
5、ttheproposedmethodhassolvedthenonconvergentproblemresultedfromtraditionalmethods.comparingwithparticleswarmoptimization,themethodgiveninthispapershowsuptotwotimesfasterconvergencerate,andforsomesubjects,thenewmethodshowstentimeshigherprecision.itisdifficulttoestima
6、tetheparametersofsoftwarereliabilitymodels,sincemostofthemarenon.linearmodels.themostwidelyusedmethodsforparametersestimatingofsoftwarereliabilitymodelshavebeensummarized,andanewapproachbasedonantcolonyalgorithmwasproposed.theexperimentswiththreetypicalmodels,g.omo
7、del,weibullmodelandm.omodel,showthatthisalgorithmdemonstratesgoodapplicability.andtheresultsdemonstratethattheproposedmethodhassolvedthenonconvergentproblemthatresultedfromtraditionalmethods.comparedwithparticleswarmoptimization(pso),themethodgiveninthispapershowsu
8、ptotwotimesfasterconvergencerate,andforsomesubjects,thenewmethodshowstentimeshigherprecision.keywords:antcolonyalgorithm;softwarereliabilitymode