基于随机Petri网的软件可靠性分析研究
|
摘要
随着软件系统应用范围的日趋广泛,我们的日常生活越来越依赖应用软件所提供的服务,目前软件已经应用于航空航天、核电技术、军事、电子机械、金融等许多重要的特殊工程领域,在这些应用领域里,一个小的软件故障可能会导致灾难性的生命和财产损失。同时随着近年来软件规模越来越大,软件系统的复杂性也急剧地增加,解决软件可靠性问题变得越来越迫切。
软件可靠性分析主要分为软件可靠性建模和可靠性指标分析。其中软件可靠性建模在可靠性分析起着核心的作用,目前所提出的模型大部分是基于故障树进行分析,不能全面地描述系统的动态行为,因而存在一定的应用范围和条件的局限性,不能很好地适应复杂多变的应用环境的需求;在可靠性指标分析方面,目前的计算方法大多只能针对失效的稳定状态进行可靠性分析,因此无法掌握整个系统的可靠度动态的变化趋势;此外,大多数软件失效是随机的、间歇的,软件可靠性模型在形式上符合随机过程的特性,这使得在可靠性分析中,时间比其他的特性更加重要。
本文首先把随机Petri网应用于可靠性建模,随机Petri网把时间参数与变迁关联,有助于定义可靠性模型的动态行为。同时本文通过比较随机Petri网及软件可靠性技术的概念、特点及它们之间的关系,从软件故障产生与传递的表述方面着眼给出了具体的可靠性建模方法;然后,针对随机Petri网应用中常常会碰到的复杂的空间爆炸问题给出了两种解决方案;其次,引入了一种计算瞬时可靠度的方法对简化后的随机Petri网进行分析,提高了评价和预测可靠性的精确度和正确性,并且所获得的结果有助于实时监控软件运行一定时间内的可靠性变化情况,从而可以指导可靠性设计的后期开发;最后,通过实验结果验证了该可靠性分析方法的可行性及有效性。
With increasing scale of softwares, the service largely depends on applied softwares in our daily life. Some softwares now are applied to special fields, such as air industry, nuclear power technology, military industry, electron mechanism and banks. The failure of softwares will contribute to big loss of lives and money. The reliability of the software crys for being solved with the increasing scale and complexity.
The work of software reliability analysis includes modeling software reliability and analysing software indexs, and the reliability model is very important. The present models mainly analyse with the failure tree, it can not describe the whole dynamic action, so it can not adapt to the quick changing environment and is restricted; Concerning the analyzing software indexs, most of calculating methods are used to analyse the failure of steady state, it makes us can not seize changing current of reliability; Besides, most of software failure are stochastic and interim ,then reliability model is formally a stochastic process, it makes time be much more important than defining dynamic logic.
Firstly, the Stochastic Petri Net is applied to reliability model. It can help define dynamic action with relation between transition and time. Meanwhile, the detailed reliability model method is given with comparing concepts ,traits and relation between Stochastic Petri Net and software reliability; secondly, two methods are given in the paper to solve the state space explosion problem ; Thirdly, a method of calculating instantaneous reliability is introduced to analyse the simplified Stochastic Petri Net, it improves precision of model and can guide the reliability design; finally, the feasibility and validity are verified by the test.
软件可靠性分析主要分为软件可靠性建模和可靠性指标分析。其中软件可靠性建模在可靠性分析起着核心的作用,目前所提出的模型大部分是基于故障树进行分析,不能全面地描述系统的动态行为,因而存在一定的应用范围和条件的局限性,不能很好地适应复杂多变的应用环境的需求;在可靠性指标分析方面,目前的计算方法大多只能针对失效的稳定状态进行可靠性分析,因此无法掌握整个系统的可靠度动态的变化趋势;此外,大多数软件失效是随机的、间歇的,软件可靠性模型在形式上符合随机过程的特性,这使得在可靠性分析中,时间比其他的特性更加重要。
本文首先把随机Petri网应用于可靠性建模,随机Petri网把时间参数与变迁关联,有助于定义可靠性模型的动态行为。同时本文通过比较随机Petri网及软件可靠性技术的概念、特点及它们之间的关系,从软件故障产生与传递的表述方面着眼给出了具体的可靠性建模方法;然后,针对随机Petri网应用中常常会碰到的复杂的空间爆炸问题给出了两种解决方案;其次,引入了一种计算瞬时可靠度的方法对简化后的随机Petri网进行分析,提高了评价和预测可靠性的精确度和正确性,并且所获得的结果有助于实时监控软件运行一定时间内的可靠性变化情况,从而可以指导可靠性设计的后期开发;最后,通过实验结果验证了该可靠性分析方法的可行性及有效性。
With increasing scale of softwares, the service largely depends on applied softwares in our daily life. Some softwares now are applied to special fields, such as air industry, nuclear power technology, military industry, electron mechanism and banks. The failure of softwares will contribute to big loss of lives and money. The reliability of the software crys for being solved with the increasing scale and complexity.
The work of software reliability analysis includes modeling software reliability and analysing software indexs, and the reliability model is very important. The present models mainly analyse with the failure tree, it can not describe the whole dynamic action, so it can not adapt to the quick changing environment and is restricted; Concerning the analyzing software indexs, most of calculating methods are used to analyse the failure of steady state, it makes us can not seize changing current of reliability; Besides, most of software failure are stochastic and interim ,then reliability model is formally a stochastic process, it makes time be much more important than defining dynamic logic.
Firstly, the Stochastic Petri Net is applied to reliability model. It can help define dynamic action with relation between transition and time. Meanwhile, the detailed reliability model method is given with comparing concepts ,traits and relation between Stochastic Petri Net and software reliability; secondly, two methods are given in the paper to solve the state space explosion problem ; Thirdly, a method of calculating instantaneous reliability is introduced to analyse the simplified Stochastic Petri Net, it improves precision of model and can guide the reliability design; finally, the feasibility and validity are verified by the test.
引文
[1]郭波,武小悦等著.系统可靠性分析[M].长沙:国防科技大学出版社,2002
[2]金星,洪延姬等著.工程系统可靠性数值分析方法[M].北京:国防工业出版社,2002
[3]姜兴渭,宋政吉,王晓晨著.可靠性工程技术[M].哈尔滨:哈尔滨工业大学出版社,2005
[4] MichaelRLyu著.软件可靠性工程手册[M].北京:电子工业出版社,1997
[5] J.D.Musa著.软件可靠性—度量、预测和应用[M].北京:机械工业出版社,1992
[6]高社生,张玲秀.可靠性理论与工程应用[M].北京:国防工业出版社,2002
[7]刘品著.可靠性工程基础[M].北京:中国计量出版社,1995
[8]方来华,吴宗之,刘骥,康荣学,关磊.基于可靠性框图法的安全功能失效概率定量分析[J].化工自动化及仪表,2008,35(2):32-35
[9]宋华,张洪钺,王行仁.T-S模糊故障树分析方法[J].控制与决策,2005, 20(8):854-859
[10]蔡鉴明,曾峰.基于GO法的供应链可靠性分析[J].公路交通科技,2007,24(3):141-144
[11]祖培,黄卫刚.大亚湾核电站外电源系统可靠性分析中GO法的应用[J].核动力工程,2003(1):68-72.
[12] Marko C , Borut M. A Dynamic Fault Tree [J]. Reliability Engineering and System Safety , 2002 ,75 (1) :83-91.
[13]袁静,李恩有,龙勇.基于动态故障树的仿真系统可靠性建模研究[J].控制工程,2007,14(S1):107-109
[14]仉俊峰,洪炳镕,袁成军.一种基于Markov模型的星载计算机系统可靠性评测的新方式[J].2007,28(4):1053-1057
[15] John D.Musa著,韩柯译.软件可靠性工程[M].北京:机械工业出版社.2003
[16]苏春,许映秋.复杂机电产品动态可靠性建模理论与方法研究[J].中国制造业信息化,2006,35(9):24-27
[17]袁崇义.Petri网原理与应用[M].北京:电子工业出版社,2005
[18]林闯.随机Petri网和系统性能评价[M].北京:清华大学出版社,2005
[19]靳文辉,葛洪伟.优化模糊Petri网参数的一种混合智能算法[J].计算机应用与软件,2008,25(9):252-255
[20]张君一,谢里阳,王正.基于广义随机Petri网的CIMS多任务可靠性研究[J].计算机工程与应用,2007,43(33):26-29
[21]钱进,张涛.基于GOOPN的复杂关联系统可靠性分析模型[J].系统工程与电子技术,2008,30(7):1370-1374
[22] Manish Mallhotra, Kishor S Trivedi.Dependability modeling using Petri-nets[J]. IEEE Trans. on Reliability,1995,44(3):428-440
[23] Kuo Chung-Hsien, Huang Han-Pang. Failure modeling and process monitoring for flexible manufacturing systems using coored timed Petri nets[J]. IEEE Trans. on Robotics and Automation,2000,16(3):301-312
[24] Balakrishnan M. Stochastic Petri nets for reliability analysis of communication networks application with alternate routing[J].Reliability Engineering and System Safety,1995,12(2):313-318
[25]汪凯,张恒喜,王卓健.基于Petri网的复杂可修复系统可靠性仿真研究[J].电光与控制,2006,13(5):50-53
[26]李培松,刘觉夫,周娟.基于随机高级Petri网改进型网格资源调度算法[J].华东交通大学学报,2008,25(1):85-88
[27] Kiran Kumar Vemuri, Automatic Synthesis of Fault Trees for Computer-Based Systems[J]. IEEE Transactions on Reliability, 1999, 48(4): 394-402
[28] Joanne B D. Developing a Low - cost high-quality Software Tool for Dynamic Fault-tree Analysis [J]. IEEE Trans. on Reliability, 2000, 49 (1) :49-58
[29] Yangsong Rcn, Joanne Bechta Duyan. Design of Reliable Systems Using Static&Dynamic Fault Tree[J]. IEEE Tram. Reliability,1998,47(3):234-244
[30]马静,张波,辛波.马尔可夫链在冗余系统可靠度求解中的应用[J].中国惯性技术学报,2007,15(2):248-251
[31]龙士工,罗文俊.随机Petri网性能等价化简与分析方法研究[J].计算机工程,2006,3(12):27-29
[32]邹逢兴,张湘平.计算机应用系统的故障诊断与可靠性技术基础.北京:高等教育出版社,1999
[33] Cabodi G, Camurati P, Quer S. Reachability Analysis of Large Circuits Using Disjunctive Partitioning and Partial Iterative Squaring[J]. Systems Architechture, 2001, 47(2): 163-179.
[34]蒋乐天.软件缺陷及软件可靠性技术[J].计算机仿真,2004,21(2):141-144
[35] Manish Malhotra, Kishor Strived. Power-hierarchy of dependability-model types. IEEE Trans. on Reliability.1994,43(3):493-501
[36] Mei H, Chen F, Feng YD et al. ABC: An architecture based, component oriented approach to software development. Journal of Software,2003,14(4):721-732
[37]贺国光,周良生.基于可靠性单元的网络系统可靠性简化算法[J].机械设计,2008,25(7):62-65
[38] Gokhale, S S &K S, Trivedi. Reliability prediction and sensitivity analysis based on software architecture[C]. IEEE 13th International Symposium on Software Reliability Engineering, 2002:64-75.
[39] Francisco Herrera and Manuel Lozano. Gradual Distributed Real-Coded Genetic Algorithms[J]. Transactions on Evolutionary Computation, 2000,4(1):43-63
[40]詹惠琴,古军,习友宝.正态分布随机Petri网的串并行机构化简[J].电子科技大学学报,2008,37(3):426-429
[41]李妍琛.利用随机Petri网分析软件可靠性[J].计算机应用与软件,2009,26(8):132-136
[42] DUGAN J B. Modular techniques for dynamic fault tree analysis[J ] .Proceedings of the 1992 Annual Reliability and Maintainability Symposium , 1992 , 10 (3) : 56 - 64.
[43] Gianf ranco C ,Reinhard G,Christoph L ,et al. A Characterization of theStochastic Process Underlying a Stochastic Petri Net [J ]. IEEE Trans . on Software Engineering ,1994 ,20 (7) :506-515
[44]朱磊,杨丹,吴映波.基于BP神经网络的软件可靠性模型选择[J].计算机工程与设计,2007,28(17):4091-4094
[45]陈俊文,谷建华,张凡,董云卫.一种改进的基于构架的软件可靠性模型[J].计算机工程与应用,2009,45(33):60-63
[2]金星,洪延姬等著.工程系统可靠性数值分析方法[M].北京:国防工业出版社,2002
[3]姜兴渭,宋政吉,王晓晨著.可靠性工程技术[M].哈尔滨:哈尔滨工业大学出版社,2005
[4] MichaelRLyu著.软件可靠性工程手册[M].北京:电子工业出版社,1997
[5] J.D.Musa著.软件可靠性—度量、预测和应用[M].北京:机械工业出版社,1992
[6]高社生,张玲秀.可靠性理论与工程应用[M].北京:国防工业出版社,2002
[7]刘品著.可靠性工程基础[M].北京:中国计量出版社,1995
[8]方来华,吴宗之,刘骥,康荣学,关磊.基于可靠性框图法的安全功能失效概率定量分析[J].化工自动化及仪表,2008,35(2):32-35
[9]宋华,张洪钺,王行仁.T-S模糊故障树分析方法[J].控制与决策,2005, 20(8):854-859
[10]蔡鉴明,曾峰.基于GO法的供应链可靠性分析[J].公路交通科技,2007,24(3):141-144
[11]祖培,黄卫刚.大亚湾核电站外电源系统可靠性分析中GO法的应用[J].核动力工程,2003(1):68-72.
[12] Marko C , Borut M. A Dynamic Fault Tree [J]. Reliability Engineering and System Safety , 2002 ,75 (1) :83-91.
[13]袁静,李恩有,龙勇.基于动态故障树的仿真系统可靠性建模研究[J].控制工程,2007,14(S1):107-109
[14]仉俊峰,洪炳镕,袁成军.一种基于Markov模型的星载计算机系统可靠性评测的新方式[J].2007,28(4):1053-1057
[15] John D.Musa著,韩柯译.软件可靠性工程[M].北京:机械工业出版社.2003
[16]苏春,许映秋.复杂机电产品动态可靠性建模理论与方法研究[J].中国制造业信息化,2006,35(9):24-27
[17]袁崇义.Petri网原理与应用[M].北京:电子工业出版社,2005
[18]林闯.随机Petri网和系统性能评价[M].北京:清华大学出版社,2005
[19]靳文辉,葛洪伟.优化模糊Petri网参数的一种混合智能算法[J].计算机应用与软件,2008,25(9):252-255
[20]张君一,谢里阳,王正.基于广义随机Petri网的CIMS多任务可靠性研究[J].计算机工程与应用,2007,43(33):26-29
[21]钱进,张涛.基于GOOPN的复杂关联系统可靠性分析模型[J].系统工程与电子技术,2008,30(7):1370-1374
[22] Manish Mallhotra, Kishor S Trivedi.Dependability modeling using Petri-nets[J]. IEEE Trans. on Reliability,1995,44(3):428-440
[23] Kuo Chung-Hsien, Huang Han-Pang. Failure modeling and process monitoring for flexible manufacturing systems using coored timed Petri nets[J]. IEEE Trans. on Robotics and Automation,2000,16(3):301-312
[24] Balakrishnan M. Stochastic Petri nets for reliability analysis of communication networks application with alternate routing[J].Reliability Engineering and System Safety,1995,12(2):313-318
[25]汪凯,张恒喜,王卓健.基于Petri网的复杂可修复系统可靠性仿真研究[J].电光与控制,2006,13(5):50-53
[26]李培松,刘觉夫,周娟.基于随机高级Petri网改进型网格资源调度算法[J].华东交通大学学报,2008,25(1):85-88
[27] Kiran Kumar Vemuri, Automatic Synthesis of Fault Trees for Computer-Based Systems[J]. IEEE Transactions on Reliability, 1999, 48(4): 394-402
[28] Joanne B D. Developing a Low - cost high-quality Software Tool for Dynamic Fault-tree Analysis [J]. IEEE Trans. on Reliability, 2000, 49 (1) :49-58
[29] Yangsong Rcn, Joanne Bechta Duyan. Design of Reliable Systems Using Static&Dynamic Fault Tree[J]. IEEE Tram. Reliability,1998,47(3):234-244
[30]马静,张波,辛波.马尔可夫链在冗余系统可靠度求解中的应用[J].中国惯性技术学报,2007,15(2):248-251
[31]龙士工,罗文俊.随机Petri网性能等价化简与分析方法研究[J].计算机工程,2006,3(12):27-29
[32]邹逢兴,张湘平.计算机应用系统的故障诊断与可靠性技术基础.北京:高等教育出版社,1999
[33] Cabodi G, Camurati P, Quer S. Reachability Analysis of Large Circuits Using Disjunctive Partitioning and Partial Iterative Squaring[J]. Systems Architechture, 2001, 47(2): 163-179.
[34]蒋乐天.软件缺陷及软件可靠性技术[J].计算机仿真,2004,21(2):141-144
[35] Manish Malhotra, Kishor Strived. Power-hierarchy of dependability-model types. IEEE Trans. on Reliability.1994,43(3):493-501
[36] Mei H, Chen F, Feng YD et al. ABC: An architecture based, component oriented approach to software development. Journal of Software,2003,14(4):721-732
[37]贺国光,周良生.基于可靠性单元的网络系统可靠性简化算法[J].机械设计,2008,25(7):62-65
[38] Gokhale, S S &K S, Trivedi. Reliability prediction and sensitivity analysis based on software architecture[C]. IEEE 13th International Symposium on Software Reliability Engineering, 2002:64-75.
[39] Francisco Herrera and Manuel Lozano. Gradual Distributed Real-Coded Genetic Algorithms[J]. Transactions on Evolutionary Computation, 2000,4(1):43-63
[40]詹惠琴,古军,习友宝.正态分布随机Petri网的串并行机构化简[J].电子科技大学学报,2008,37(3):426-429
[41]李妍琛.利用随机Petri网分析软件可靠性[J].计算机应用与软件,2009,26(8):132-136
[42] DUGAN J B. Modular techniques for dynamic fault tree analysis[J ] .Proceedings of the 1992 Annual Reliability and Maintainability Symposium , 1992 , 10 (3) : 56 - 64.
[43] Gianf ranco C ,Reinhard G,Christoph L ,et al. A Characterization of theStochastic Process Underlying a Stochastic Petri Net [J ]. IEEE Trans . on Software Engineering ,1994 ,20 (7) :506-515
[44]朱磊,杨丹,吴映波.基于BP神经网络的软件可靠性模型选择[J].计算机工程与设计,2007,28(17):4091-4094
[45]陈俊文,谷建华,张凡,董云卫.一种改进的基于构架的软件可靠性模型[J].计算机工程与应用,2009,45(33):60-63