云模型在系统可靠性中的应用研究
|
摘要
可靠性从上个世纪三四十年代就受到了人们的关注,其主要原因就是当时军用产品频频出现故障,以至于人们不得不深刻的反思应该怎样对产品的可靠性进行保证。从那以后,人们便开始了对可靠性的研究。许多国家都相继建立了自己的可靠性研究机构,将可靠性用到了生活的方方面面,对产品质量的提高和生活安全的保障都起到了很好的促进作用。
总体说来,可靠性的发展大体上分为三个阶段。第一个阶段就是上三四十年代,也就是可靠性的发展的初期阶段,其主要的原因就是当时军用的电子设备经常出现故障,引发了一系列对可靠性思考,使得可靠性的研究开始进入人们的视野。第二个阶段是上个世纪五六十年代,也就是可靠性的中期发展阶段,这个时期主要的发展体现在以下三个方面：一、研究人员将概率论与数理统计的相关知识用到了可靠性评估中来,并且对许多的可靠性问题进行了深入的研究。二、许多国家都相继建立了自己的可靠性组织,使得可靠性的研究进入了专业性的轨道。但这时的可靠性研究主要偏重于民用产品的可靠性研究,带来了很好的社会效益。三、可靠性被应用到了生产和生活的方方面面,例如,电气,化工,冶金,建筑,食品,通信,医疗等多个方面。第三个阶段就是上个世纪七十年代以后,这时的主要成就集中在两个方面。一、许多国家都颁布了自己的可靠性标准,也出现了一些国际上通用的可靠性标准。二、形成了一系列比较成熟的可靠性评估方法。其中包括,故障树分析(FTA),事件树分析(ETA),故障模式影响及可靠性分析(FMECA)等。它们都在实际生活中得到了广泛的应用,也收到了很好的效果。
虽然上面提出了很多的可靠性评估方法,但这些方法只能很好的解决在一个或者多个工作条件下的可靠性的评估问题,对于那些工作条件变化,或者工作条件用定性语言描述的情况下的可靠性评估,就不能适用。为了解决此类问题,宋远骏,李德毅等人提出了基于云模型的可靠性评估方法,解决了此问题。
在生活中,我们经常会遇到一些含有不确定性的语言值,例如“20公里左右”,“青年人”等,在可靠性评估中,也经常遇到这样的语言值,例如“寒冷”,“电压不稳”,等。为了更好的描述这种不确定性,李德毅院土于1995年提出了一种定性与定量的转换模型—云模型。云模型可以将含有不确定性的定性的语言值进行量化,从而更直观的表达各种不确定性。宋远骏,李德毅等提出了基于云模型的可靠性评估方法。此种方法的主要的特点是将产品的工作条件转换为工作条件云,并建立产品的环境适应度云,然后将不同的工作条件代入相应的环境适应度云,得到在不同的工作条件下的环境适应度,最后计算出其可靠性。这样的计算是一个多次随机的过程,多次随机的结果反映了产品可靠性的总体变化情况。在采用云模型进行可靠性评估时,对其评估结果有两点要注意：一、其可靠性评估结果不是一个确定的值,基本上都是在一定的范围内波动,其主要的原因就是描述工作条件的语言值含有不确定性。二、随着时间推移,其波动范围是不一样的。一股情况下,开始时系统可靠性较高,其波动的范围也很小,工作时间越长,其可靠性会变低,其波动的范围也会变大。实验证明,这种基于云模型的系统可靠性评估方法很好的解决了在工作条件可变或者工作条件用定性语言描述的情况的可靠性评估问题。
但是这种基于云模型的可靠性评估方法,也有几个问题值得去进行深入的研究：
(1)上面介绍的基于云模型的可靠性评估方法是针对单个产品或者部件的,那么对于具有更加复杂的结构的系统(如串联,并联,串并混联)如何使用云模型进行可靠性评估呢?
(2)从实际生活中我们知道,不同的环境因素对产品可靠性的大小的影响是不一样的,同一因素对不同的产品的可靠性的影响也是不一样的,那么,在使用云模型进行可靠性评估时,如何将这些因素考虑进去,使得评估的结果更加符合实际?
针对上面这两个问题,我们提出了基于云模型的串并联系统的可靠性评估方法和加权云模型,其主要的思想如下：
(1)基于云模型的串并联系统可靠性评估方法。由于传统的串并联系统的可靠性评估方法可以在工作条件恒定的情况下进行定量的计算,考虑到在云模型的可靠性评估过程中,每一次随机的随机的可靠性评估实际上也就是一次定量的评估,因此,在计算时可以采取传统的串联系统的可靠性评估方法,也就是说,在大体的云模型随机实现的基础之上,随机之后的具体的计算是通过传统的串联系统的可靠性评估方式进行的,实际上是一种将云模型和传统的可靠性评估方式相结合的可靠性评估方式。
(2)加权云模型。为了刻画不同的环境因素对可靠性的影响,最直接的方法就是对不同的环境因素赋予不同的权重。对那些对可靠性影响很大的环境因素赋予大的权重,对于那些对环境因素影响不大的因素,赋予小的权重。这样使得不同的环境因素的变化对产品的可靠性的影响是不一样。并且我们将加权云模型应用于单个部件和串并联系统的可靠性评估,这种方法同样也将传统的可靠性评估方法和云模型结合起来,使得评价结果更加符合实际情况。
文中通过实验证明了我们提出这两种基于云模型的可靠性评方式的有效性和实用性,为可靠性的评估提供了一种新的思路。
Reliability has been paid attention to from last century in 1930's to 1940's. The main reason is that many military products broken down frequently. So people start to think about the reliability of products. From then on, many people start to research the methods which can be used for reliability evaluation. Many countries start to build their own reliability research organization, and reliability evaluation has been used in many fields in our real life. It promotes the quality of products, and ensures the safety of our life.
Generally speaking, the development of reliability can be divided into three stages. First stage which is from 1930's to 1940's in last century. Many problems arise from military products. This brings reliability evaluation to our vision. Second stage which is from 1950's to 1960's. The accomplishment in this stage can be classified in to three categories. First, researchers combine the probability theory and mathematical statistics with reliability. They also conducted in-depth research about some reliability evaluation problems. Second, many countries have build their reliability research organizations. It makes the reliability research to be very professional. The reliability research in this stage is focused on the civil products. It brings great profit to our society. Third, reliability has been used in many fields, such as electrical science, chemical engineering, metallurgy, communication, medical treatment, etc. Third stage which is from 1970's, the accomplishment in this stage can be classified in to two categories. First, many countries have published their own criterion of reliability. There are also many other international criterions of reliability. Second, Many reliability evaluation methods appeared in this stage, which include Fault Tree Analysis (FTA), Event Tree Analysis (ETA), Failure Modes Effects and Criticality Analysis (FMECA), etc. These methods have been used in our life successfully.
Though these methods presented above are very practical, there still many problems needed to be paid attention to. For instances, these traditional methods work very well in special working conditions, but it is not suitable when the working conditions are described by qualitative language values. Due to this, reliability evaluation method based on cloud model is proposed by Song Yunjun, Li Deyi etc.
In our real life, we always use many words which contain uncertainty such as "about 20 kilometers", "young man", etc. In order to better description the uncertainty, Cloud model is proposed by Li Deyi in 1995, which can be used for the transformation between the qualitative language values and quantitative values. Song Yuanjun, Li Deyi, etc. proposed the reliability evaluation method based on cloud model. The basic flow of reliability evaluation using cloud model can be divided into four steps. First, construct the working condition cloud and suitability cloud. Second, take different working conditions which are produced by working condition cloud as parameters input into the suitability cloud model, and then compute the suitability degree and MTTF of products under different working conditions. Third, compute the reliability of products under different working conditions. Forth, Many times of stochastic implements reflect the reliability of products. There are still two points need further research:
(1) The reliability evaluation method presented above is only used for single product and component, what about some other complex systems, such as series systems, parallel systems?
(2) According to our experiences, different environmental factors may have different effect on reliability; also same environmental factors may have different effect on the same products. How to reflect this when using the reliability evaluation method based on cloud model?
In order to solve these two problems, we proposed the following two methods:reliability evaluation method for series-parallel systems based on cloud model, weighted cloud model. The main idea is as follows:
(1) Reliability evaluation method for series-parallel system. Because traditional reliability evaluation method for series-parallel system works very well under specific working conditions or constant working conditions. Considering that each time of stochastic implement based on cloud model can be viewed as under constant or specific working conditions. So according to the structure of system, the reliability of system can be obtained by many times of stochastic implements which reliability is obtained by the traditional reliability evaluation method. In fact, this method is a hybrid method based on cloud model and traditional reliability evaluation. (2) Weighted cloud model. In order to describe different effects caused by the different environmental factors, different weight values are assigned by to different environmental factors. The environmental factors which have big weighted values may have big effect on the reliability of products; otherwise, the environmental factors may have small effects on the reliability. This makes the reliability evaluation results closer to the real situation.
Experimental results prove that these two methods are very practical and useful. It may have border and potential applications in our real life.
总体说来,可靠性的发展大体上分为三个阶段。第一个阶段就是上三四十年代,也就是可靠性的发展的初期阶段,其主要的原因就是当时军用的电子设备经常出现故障,引发了一系列对可靠性思考,使得可靠性的研究开始进入人们的视野。第二个阶段是上个世纪五六十年代,也就是可靠性的中期发展阶段,这个时期主要的发展体现在以下三个方面：一、研究人员将概率论与数理统计的相关知识用到了可靠性评估中来,并且对许多的可靠性问题进行了深入的研究。二、许多国家都相继建立了自己的可靠性组织,使得可靠性的研究进入了专业性的轨道。但这时的可靠性研究主要偏重于民用产品的可靠性研究,带来了很好的社会效益。三、可靠性被应用到了生产和生活的方方面面,例如,电气,化工,冶金,建筑,食品,通信,医疗等多个方面。第三个阶段就是上个世纪七十年代以后,这时的主要成就集中在两个方面。一、许多国家都颁布了自己的可靠性标准,也出现了一些国际上通用的可靠性标准。二、形成了一系列比较成熟的可靠性评估方法。其中包括,故障树分析(FTA),事件树分析(ETA),故障模式影响及可靠性分析(FMECA)等。它们都在实际生活中得到了广泛的应用,也收到了很好的效果。
虽然上面提出了很多的可靠性评估方法,但这些方法只能很好的解决在一个或者多个工作条件下的可靠性的评估问题,对于那些工作条件变化,或者工作条件用定性语言描述的情况下的可靠性评估,就不能适用。为了解决此类问题,宋远骏,李德毅等人提出了基于云模型的可靠性评估方法,解决了此问题。
在生活中,我们经常会遇到一些含有不确定性的语言值,例如“20公里左右”,“青年人”等,在可靠性评估中,也经常遇到这样的语言值,例如“寒冷”,“电压不稳”,等。为了更好的描述这种不确定性,李德毅院土于1995年提出了一种定性与定量的转换模型—云模型。云模型可以将含有不确定性的定性的语言值进行量化,从而更直观的表达各种不确定性。宋远骏,李德毅等提出了基于云模型的可靠性评估方法。此种方法的主要的特点是将产品的工作条件转换为工作条件云,并建立产品的环境适应度云,然后将不同的工作条件代入相应的环境适应度云,得到在不同的工作条件下的环境适应度,最后计算出其可靠性。这样的计算是一个多次随机的过程,多次随机的结果反映了产品可靠性的总体变化情况。在采用云模型进行可靠性评估时,对其评估结果有两点要注意：一、其可靠性评估结果不是一个确定的值,基本上都是在一定的范围内波动,其主要的原因就是描述工作条件的语言值含有不确定性。二、随着时间推移,其波动范围是不一样的。一股情况下,开始时系统可靠性较高,其波动的范围也很小,工作时间越长,其可靠性会变低,其波动的范围也会变大。实验证明,这种基于云模型的系统可靠性评估方法很好的解决了在工作条件可变或者工作条件用定性语言描述的情况的可靠性评估问题。
但是这种基于云模型的可靠性评估方法,也有几个问题值得去进行深入的研究：
(1)上面介绍的基于云模型的可靠性评估方法是针对单个产品或者部件的,那么对于具有更加复杂的结构的系统(如串联,并联,串并混联)如何使用云模型进行可靠性评估呢?
(2)从实际生活中我们知道,不同的环境因素对产品可靠性的大小的影响是不一样的,同一因素对不同的产品的可靠性的影响也是不一样的,那么,在使用云模型进行可靠性评估时,如何将这些因素考虑进去,使得评估的结果更加符合实际?
针对上面这两个问题,我们提出了基于云模型的串并联系统的可靠性评估方法和加权云模型,其主要的思想如下：
(1)基于云模型的串并联系统可靠性评估方法。由于传统的串并联系统的可靠性评估方法可以在工作条件恒定的情况下进行定量的计算,考虑到在云模型的可靠性评估过程中,每一次随机的随机的可靠性评估实际上也就是一次定量的评估,因此,在计算时可以采取传统的串联系统的可靠性评估方法,也就是说,在大体的云模型随机实现的基础之上,随机之后的具体的计算是通过传统的串联系统的可靠性评估方式进行的,实际上是一种将云模型和传统的可靠性评估方式相结合的可靠性评估方式。
(2)加权云模型。为了刻画不同的环境因素对可靠性的影响,最直接的方法就是对不同的环境因素赋予不同的权重。对那些对可靠性影响很大的环境因素赋予大的权重,对于那些对环境因素影响不大的因素,赋予小的权重。这样使得不同的环境因素的变化对产品的可靠性的影响是不一样。并且我们将加权云模型应用于单个部件和串并联系统的可靠性评估,这种方法同样也将传统的可靠性评估方法和云模型结合起来,使得评价结果更加符合实际情况。
文中通过实验证明了我们提出这两种基于云模型的可靠性评方式的有效性和实用性,为可靠性的评估提供了一种新的思路。
Reliability has been paid attention to from last century in 1930's to 1940's. The main reason is that many military products broken down frequently. So people start to think about the reliability of products. From then on, many people start to research the methods which can be used for reliability evaluation. Many countries start to build their own reliability research organization, and reliability evaluation has been used in many fields in our real life. It promotes the quality of products, and ensures the safety of our life.
Generally speaking, the development of reliability can be divided into three stages. First stage which is from 1930's to 1940's in last century. Many problems arise from military products. This brings reliability evaluation to our vision. Second stage which is from 1950's to 1960's. The accomplishment in this stage can be classified in to three categories. First, researchers combine the probability theory and mathematical statistics with reliability. They also conducted in-depth research about some reliability evaluation problems. Second, many countries have build their reliability research organizations. It makes the reliability research to be very professional. The reliability research in this stage is focused on the civil products. It brings great profit to our society. Third, reliability has been used in many fields, such as electrical science, chemical engineering, metallurgy, communication, medical treatment, etc. Third stage which is from 1970's, the accomplishment in this stage can be classified in to two categories. First, many countries have published their own criterion of reliability. There are also many other international criterions of reliability. Second, Many reliability evaluation methods appeared in this stage, which include Fault Tree Analysis (FTA), Event Tree Analysis (ETA), Failure Modes Effects and Criticality Analysis (FMECA), etc. These methods have been used in our life successfully.
Though these methods presented above are very practical, there still many problems needed to be paid attention to. For instances, these traditional methods work very well in special working conditions, but it is not suitable when the working conditions are described by qualitative language values. Due to this, reliability evaluation method based on cloud model is proposed by Song Yunjun, Li Deyi etc.
In our real life, we always use many words which contain uncertainty such as "about 20 kilometers", "young man", etc. In order to better description the uncertainty, Cloud model is proposed by Li Deyi in 1995, which can be used for the transformation between the qualitative language values and quantitative values. Song Yuanjun, Li Deyi, etc. proposed the reliability evaluation method based on cloud model. The basic flow of reliability evaluation using cloud model can be divided into four steps. First, construct the working condition cloud and suitability cloud. Second, take different working conditions which are produced by working condition cloud as parameters input into the suitability cloud model, and then compute the suitability degree and MTTF of products under different working conditions. Third, compute the reliability of products under different working conditions. Forth, Many times of stochastic implements reflect the reliability of products. There are still two points need further research:
(1) The reliability evaluation method presented above is only used for single product and component, what about some other complex systems, such as series systems, parallel systems?
(2) According to our experiences, different environmental factors may have different effect on reliability; also same environmental factors may have different effect on the same products. How to reflect this when using the reliability evaluation method based on cloud model?
In order to solve these two problems, we proposed the following two methods:reliability evaluation method for series-parallel systems based on cloud model, weighted cloud model. The main idea is as follows:
(1) Reliability evaluation method for series-parallel system. Because traditional reliability evaluation method for series-parallel system works very well under specific working conditions or constant working conditions. Considering that each time of stochastic implement based on cloud model can be viewed as under constant or specific working conditions. So according to the structure of system, the reliability of system can be obtained by many times of stochastic implements which reliability is obtained by the traditional reliability evaluation method. In fact, this method is a hybrid method based on cloud model and traditional reliability evaluation. (2) Weighted cloud model. In order to describe different effects caused by the different environmental factors, different weight values are assigned by to different environmental factors. The environmental factors which have big weighted values may have big effect on the reliability of products; otherwise, the environmental factors may have small effects on the reliability. This makes the reliability evaluation results closer to the real situation.
Experimental results prove that these two methods are very practical and useful. It may have border and potential applications in our real life.
引文
[1]陈晓彤,赵廷弟,王云飞,等.可靠性实用指南[M],北京：北京航空航天大学出版社,2005.
[3]Lee Yun-Seong, Kim Hyung-Chul, Cha Jun-Min, etc. A new method for FMECA using expert system and fuzzy,theory. In Proc. of 9th Conference on Environment and Electrical Engineering,293-296,2010.
[4]Guo Linhan, Xiao Boping, Shi Rongde, etc. Corrective maintenance task ascertain method research based on FMECA. In proc. of 8th International Conference on Reliability, Maintainability and Satety,665-669,2009.
[5]Mili, A, Siadat, A,Hubac,S. etc. Dynamic management of detected factory events and estimated risks using FMECA. In Proc. of 4th IEEE International Conference on Management of Innovation and Technology,1204-1209,2008.
[6]Liu Wensheng, Guo Liwen, ZhuMing. Bayesian network based on FTA for safety evaluation on coalmine haulage system. In Proc. of First international Information Computing and Applications.143-149,2010.
[7]Zhou Liming,CAi Guoqiang, Yang jianwei, etc. Monte-Carlo simulation based on FTA in reliability analysis of door system.713-717,2010.
[8]Yang Shunkun, Lu Mingyan, Liu Bin,etc. A fault diagnosis model for embedded software based on FMECA/FTA and Bayesian Network. In Proc. of 8th International Conference on Reliability,Maintainability and safety,778-782,2009.
[9]Kohda T. Inoue K. Fault-tree analysis considering latency of basic events, In Proc. of the Annual Reliability and Maintainability Symposium,32-35,2001.
[10]Metzroth Kyle, Denning Richard, Aldemir Tunc. Dynamic event tree analysis. In Proc. of International Congress on Advances in Nuclear Power Plants,979-986,2010.
[11]Ahmadi Alireza, Soderholm Peter. Assessment of operational consequences of aircraft failures: using event tree analysis. In Proc. of IEEE Aerospace Conference,2008.
[12]李凌,徐伟.威布尔产品加速寿命试验的可靠性分析[J],2010,32(7)：1544-1548.
[13]丁湛,黄双华.基于威布尔分布的可靠性寿命分布的模型的建立[J],电子测量技术2007,(3)：34-35.
[14]秦明,巫世晶,彭潇,等.一种服从威布尔分布的可靠性评估方法[J],武汉大学学报(工学版),2008,41(6)：100-102.
[15]Yi-Chih Hsieh, Ta-Cheng Chen, Dennis L. Bricker, Genetic algorithms for reliability design problems [J], Microelectronics Reliability,38 (1998):1599-1605.
[16]余东.遗传算法在系统可靠性优化中的应用j研究[J],武汉科技大学学报,2004,27(4)：427-428.
[17]蔡林峰,谭观音.基于遗传算法的信息系统可靠性优化设计[J],计算机工程与设计,2006,27(14)：2579-2580
[18]王建成.基于遗传算法的系统可靠性优化[J]装备指挥技术学院学报,2005,16(4)
[19]程世娟,卢伟,何平.串并联系统可靠性优化的蚁群算法[J],计算机应用与软件,2010,27(1)：38-39.
[20]程世娟,卢伟,何平.蚁群算法在复杂系统可靠性优化中的应用[J],工程设计学报,2009,16(3)：178-181.
[21]高尚,杨静宇,吴小俊,等,可靠性优化的蚁群算法[J],计算机应用与软件,2004,21(12)：94-96.
[22]高尚,杨静宇.群智能算法及其应用[M].北京：中国水利水电出版社,2006.
[23]王正初,李薇薇.基于粒子群算法的可靠性优化[J],台洲学院学报,2006,28(6)：29-32.
[24]高尚,杨静宇.可靠性优化的一种新的算法[J],工程设计学报,2006,13(2)：74-77.
[25]高尚,基于模拟退火算法的可靠性优化[J],上海航天,2002,12(2)：21-23.
[26]高尚.系统可靠性优化方法[J].上海航天,2001,18(3)：36-40.
[27]李德毅,孟海军,史雪梅.隶属云和隶属云发生器[J],计算机研究与发展,1995,32(6)：16-21.
[28]宋远骏,李德毅,杨孝宗等.电子产品可靠性的云模型评价方法[J],电子学报,2000,28(12)：74-76.
[29]宋远骏,杨孝宗,李德毅等考虑环境因素的计算机可靠性云模型评价[J],计算机研究与发展,2001,38(5)：631-635.
[30]宋政吉,王慧,王立.相容非概率信息的可靠性理论与方法研究[J],航天器工程,2007,16(4)：29-35.
[31]宋笔锋,等.飞行器可靠性工程[M].北京：北京航空航天大学出版社,2006.
[32]赵宇,杨军,等.可靠性数据分析教程[M].北京：北京航空航天大学出版社,2009.
[33]周正伐,顾长鸿,等.航天可靠性工程[M].北京：中国宇航出版社,2007.
[34]冯静,孙权,等.装备可靠性与综合保障[M],北京：国防科技大学出版社,2008.
[35]李德毅,杜鹢,不确定性人工智能[M].北京：国防工业出版社,2005.
[36]李德毅,刘常昱,杜鹚等.不确定性人工智能[J],软件学报,2004,15(11)：1583-1593.
[37]Zhu Yunfang, Dai Chaohua, Chen Weirong, Lin Jianhui.Adaptive probabilities of crossover and mutation in genetic algorithms based on cloud generators [J], Journal of Computational Information Systems 2005,1 (4):671-678.
[38]戴朝华,朱云芳,陈维荣等,云遗传算法及其应用[J],电子学报,2007,35(7)：1419-1424
[39]张光卫,何锐,刘禹等.基于云模型的进化算法[J],计算机学报,2008,31(7)：1083-1091
[40]戴朝华,朱云芳,陈维荣.云遗传算法[J],西南交通大学学报,2006,41(6)：729-732
[41]Pin Lv, Lin Yuan, Jinfang Zhang Cloud theory-based simulated annealing algorithm and application [J], Engineering Applications of Artifcial Intelligence 22 (2009) 742-749.
[42]Xiaolan Wu, Bo Cheng, Jianbo Cao et.al. Particle Swarm Optimization with Normal Cloud Mutation. Proceedings of the 7th World Congress on Intelligent Control and Automation WCICA,2008:2828-2832.
[43]刘桂花.基于云模型的关联规则的研究[D],山东师范大学硕士学位论文,2007.
[44]冯朝一.云理论在数据挖掘中的应用研究[D],广西大学硕士学位论文,2007.
[45]Li De-yi, Di Kai-chang, Li De-ren, Shi Xuemei. Mining Association Rules with Linguistic Cloud Models [J] Journal of Software,2000,11(2):143—158
[46]李兴生.基于云模型和数据场的分类和聚类挖掘研究[D].中国人民解放军理工大学硕士学位论文,2003.16-19.
[47]刘继,邓贵仕,那日萨.贝叶斯反馈云模型的分析与设计[J],系统工程理论与实践,2008,(7)：139-143.
[48]赵卫伟,李德毅.基于云模型的入侵检测方法[J],计算机工程与应用,2003,39(26)：158-160.
[49]张飞舟,范跃祖,沈程智,李德毅.基于隶属云发生器的智能控制[J],航空学报,1999,20(1)：89-92
[50]杜湘瑜,尹全军,黄柯棣等,基于云模型的定性定量转换方法及其应用[J],系统工程与电子技术,2008,30(4)：773-776
[51]柳炳祥,李海林,杨丽彬.云决策分析方法[J],控制与决策,2009,24(6).
[52]刘常昱,李德毅,杜鹢.正态云模型的统计分析[J],信息与控制,2005,34(2)：236-248
[53]李德毅,刘常昱.论正态云模型的普适性[J],中国工程科学,2004,6(8)：29-31.
[3]Lee Yun-Seong, Kim Hyung-Chul, Cha Jun-Min, etc. A new method for FMECA using expert system and fuzzy,theory. In Proc. of 9th Conference on Environment and Electrical Engineering,293-296,2010.
[4]Guo Linhan, Xiao Boping, Shi Rongde, etc. Corrective maintenance task ascertain method research based on FMECA. In proc. of 8th International Conference on Reliability, Maintainability and Satety,665-669,2009.
[5]Mili, A, Siadat, A,Hubac,S. etc. Dynamic management of detected factory events and estimated risks using FMECA. In Proc. of 4th IEEE International Conference on Management of Innovation and Technology,1204-1209,2008.
[6]Liu Wensheng, Guo Liwen, ZhuMing. Bayesian network based on FTA for safety evaluation on coalmine haulage system. In Proc. of First international Information Computing and Applications.143-149,2010.
[7]Zhou Liming,CAi Guoqiang, Yang jianwei, etc. Monte-Carlo simulation based on FTA in reliability analysis of door system.713-717,2010.
[8]Yang Shunkun, Lu Mingyan, Liu Bin,etc. A fault diagnosis model for embedded software based on FMECA/FTA and Bayesian Network. In Proc. of 8th International Conference on Reliability,Maintainability and safety,778-782,2009.
[9]Kohda T. Inoue K. Fault-tree analysis considering latency of basic events, In Proc. of the Annual Reliability and Maintainability Symposium,32-35,2001.
[10]Metzroth Kyle, Denning Richard, Aldemir Tunc. Dynamic event tree analysis. In Proc. of International Congress on Advances in Nuclear Power Plants,979-986,2010.
[11]Ahmadi Alireza, Soderholm Peter. Assessment of operational consequences of aircraft failures: using event tree analysis. In Proc. of IEEE Aerospace Conference,2008.
[12]李凌,徐伟.威布尔产品加速寿命试验的可靠性分析[J],2010,32(7)：1544-1548.
[13]丁湛,黄双华.基于威布尔分布的可靠性寿命分布的模型的建立[J],电子测量技术2007,(3)：34-35.
[14]秦明,巫世晶,彭潇,等.一种服从威布尔分布的可靠性评估方法[J],武汉大学学报(工学版),2008,41(6)：100-102.
[15]Yi-Chih Hsieh, Ta-Cheng Chen, Dennis L. Bricker, Genetic algorithms for reliability design problems [J], Microelectronics Reliability,38 (1998):1599-1605.
[16]余东.遗传算法在系统可靠性优化中的应用j研究[J],武汉科技大学学报,2004,27(4)：427-428.
[17]蔡林峰,谭观音.基于遗传算法的信息系统可靠性优化设计[J],计算机工程与设计,2006,27(14)：2579-2580
[18]王建成.基于遗传算法的系统可靠性优化[J]装备指挥技术学院学报,2005,16(4)
[19]程世娟,卢伟,何平.串并联系统可靠性优化的蚁群算法[J],计算机应用与软件,2010,27(1)：38-39.
[20]程世娟,卢伟,何平.蚁群算法在复杂系统可靠性优化中的应用[J],工程设计学报,2009,16(3)：178-181.
[21]高尚,杨静宇,吴小俊,等,可靠性优化的蚁群算法[J],计算机应用与软件,2004,21(12)：94-96.
[22]高尚,杨静宇.群智能算法及其应用[M].北京：中国水利水电出版社,2006.
[23]王正初,李薇薇.基于粒子群算法的可靠性优化[J],台洲学院学报,2006,28(6)：29-32.
[24]高尚,杨静宇.可靠性优化的一种新的算法[J],工程设计学报,2006,13(2)：74-77.
[25]高尚,基于模拟退火算法的可靠性优化[J],上海航天,2002,12(2)：21-23.
[26]高尚.系统可靠性优化方法[J].上海航天,2001,18(3)：36-40.
[27]李德毅,孟海军,史雪梅.隶属云和隶属云发生器[J],计算机研究与发展,1995,32(6)：16-21.
[28]宋远骏,李德毅,杨孝宗等.电子产品可靠性的云模型评价方法[J],电子学报,2000,28(12)：74-76.
[29]宋远骏,杨孝宗,李德毅等考虑环境因素的计算机可靠性云模型评价[J],计算机研究与发展,2001,38(5)：631-635.
[30]宋政吉,王慧,王立.相容非概率信息的可靠性理论与方法研究[J],航天器工程,2007,16(4)：29-35.
[31]宋笔锋,等.飞行器可靠性工程[M].北京：北京航空航天大学出版社,2006.
[32]赵宇,杨军,等.可靠性数据分析教程[M].北京：北京航空航天大学出版社,2009.
[33]周正伐,顾长鸿,等.航天可靠性工程[M].北京：中国宇航出版社,2007.
[34]冯静,孙权,等.装备可靠性与综合保障[M],北京：国防科技大学出版社,2008.
[35]李德毅,杜鹢,不确定性人工智能[M].北京：国防工业出版社,2005.
[36]李德毅,刘常昱,杜鹚等.不确定性人工智能[J],软件学报,2004,15(11)：1583-1593.
[37]Zhu Yunfang, Dai Chaohua, Chen Weirong, Lin Jianhui.Adaptive probabilities of crossover and mutation in genetic algorithms based on cloud generators [J], Journal of Computational Information Systems 2005,1 (4):671-678.
[38]戴朝华,朱云芳,陈维荣等,云遗传算法及其应用[J],电子学报,2007,35(7)：1419-1424
[39]张光卫,何锐,刘禹等.基于云模型的进化算法[J],计算机学报,2008,31(7)：1083-1091
[40]戴朝华,朱云芳,陈维荣.云遗传算法[J],西南交通大学学报,2006,41(6)：729-732
[41]Pin Lv, Lin Yuan, Jinfang Zhang Cloud theory-based simulated annealing algorithm and application [J], Engineering Applications of Artifcial Intelligence 22 (2009) 742-749.
[42]Xiaolan Wu, Bo Cheng, Jianbo Cao et.al. Particle Swarm Optimization with Normal Cloud Mutation. Proceedings of the 7th World Congress on Intelligent Control and Automation WCICA,2008:2828-2832.
[43]刘桂花.基于云模型的关联规则的研究[D],山东师范大学硕士学位论文,2007.
[44]冯朝一.云理论在数据挖掘中的应用研究[D],广西大学硕士学位论文,2007.
[45]Li De-yi, Di Kai-chang, Li De-ren, Shi Xuemei. Mining Association Rules with Linguistic Cloud Models [J] Journal of Software,2000,11(2):143—158
[46]李兴生.基于云模型和数据场的分类和聚类挖掘研究[D].中国人民解放军理工大学硕士学位论文,2003.16-19.
[47]刘继,邓贵仕,那日萨.贝叶斯反馈云模型的分析与设计[J],系统工程理论与实践,2008,(7)：139-143.
[48]赵卫伟,李德毅.基于云模型的入侵检测方法[J],计算机工程与应用,2003,39(26)：158-160.
[49]张飞舟,范跃祖,沈程智,李德毅.基于隶属云发生器的智能控制[J],航空学报,1999,20(1)：89-92
[50]杜湘瑜,尹全军,黄柯棣等,基于云模型的定性定量转换方法及其应用[J],系统工程与电子技术,2008,30(4)：773-776
[51]柳炳祥,李海林,杨丽彬.云决策分析方法[J],控制与决策,2009,24(6).
[52]刘常昱,李德毅,杜鹢.正态云模型的统计分析[J],信息与控制,2005,34(2)：236-248
[53]李德毅,刘常昱.论正态云模型的普适性[J],中国工程科学,2004,6(8)：29-31.