周 莉,蘆 雪 娟,王 偉 華
(齊齊哈爾大學(xué) 理學(xué)院, 黑龍江 齊齊哈爾 161006 )
具有兩種修復(fù)方法的可修復(fù)系統(tǒng)解研究
周 莉*,蘆 雪 娟,王 偉 華
(齊齊哈爾大學(xué) 理學(xué)院, 黑龍江 齊齊哈爾 161006 )
將半離散算法應(yīng)用到具有兩種修復(fù)方法的可修復(fù)系統(tǒng)模型中,在[0,x0]上對(duì)其修復(fù)率進(jìn)行離散,得到了該系統(tǒng)的半離散化模型.進(jìn)一步利用泛函分析中算子半群理論將半離散后的偏微分方程轉(zhuǎn)化為抽象Cauchy問(wèn)題,即轉(zhuǎn)化為矩陣常微分方程組;再根據(jù)Trotter逼近定理證明了矩陣常微分方程組的解收斂于原方程的解.最后在故障率和修復(fù)率均為常數(shù)的前提下,利用Matlab對(duì)該系統(tǒng)的穩(wěn)定性和可靠性等進(jìn)行了數(shù)值試驗(yàn)并得到了該模型的數(shù)值解,同時(shí)給出了相應(yīng)的圖形趨勢(shì).結(jié)果表明,對(duì)具有兩種修復(fù)方法的可修復(fù)系統(tǒng)模型進(jìn)行半離散化研究,既可以為利用計(jì)算機(jī)進(jìn)一步進(jìn)行數(shù)值計(jì)算打下理論基礎(chǔ),又有助于研究和分析系統(tǒng)的可靠性.
可修復(fù)系統(tǒng);半離散化;收斂;數(shù)值計(jì)算
可修復(fù)系統(tǒng)在可靠性理論中占有重要地位,是非常重要的系統(tǒng),也是可靠性數(shù)學(xué)研究的基本問(wèn)題之一.目前,Gupur[1]、Li等[2]、Chung[3]從統(tǒng)計(jì)獨(dú)立事件和兩個(gè)狀態(tài)單調(diào)相關(guān)聯(lián)的系統(tǒng)到多個(gè)狀態(tài)相對(duì)復(fù)雜系統(tǒng)的可靠性都進(jìn)行了研究.對(duì)于一個(gè)可修復(fù)系統(tǒng)的可靠性來(lái)說(shuō),最希望的是設(shè)計(jì)出來(lái)的系統(tǒng)能長(zhǎng)時(shí)間地安全穩(wěn)定工作.對(duì)維修性來(lái)說(shuō),希望設(shè)計(jì)出來(lái)的系統(tǒng)在發(fā)生故障時(shí)能夠被快速修復(fù)好.將良好的可靠性與良好的維修性結(jié)合起來(lái),就可以保證系統(tǒng)較高的實(shí)用性.因此,在實(shí)際中為了提高系統(tǒng)的可靠性,經(jīng)常采用檢修的手段對(duì)系統(tǒng)進(jìn)行維護(hù).可修復(fù)系統(tǒng)一般由一些故障部件和一個(gè)或多個(gè)維修設(shè)備組成,維修設(shè)備對(duì)發(fā)生故障的部件進(jìn)行檢查和維修,修理后的部件可繼續(xù)執(zhí)行正常的工作.
Dhillon[4]運(yùn)用Laplace變換研究了具有兩種修復(fù)方法的可修復(fù)系統(tǒng)模型,得到了穩(wěn)態(tài)解的存在性.張玉峰等[5]證明了該系統(tǒng)動(dòng)態(tài)非負(fù)解是存在且唯一的.趙玉榮等[6]通過(guò)系統(tǒng)算子的譜點(diǎn)分析得出了其解的漸進(jìn)穩(wěn)定性及系統(tǒng)穩(wěn)態(tài)解就是系統(tǒng)算子的零本征值對(duì)應(yīng)的本征向量.本文在文獻(xiàn)[6]的基礎(chǔ)上將半離散算法[7]應(yīng)用于該系統(tǒng)模型中,對(duì)系統(tǒng)的修復(fù)率μj(x)(j=1,2)用初等階梯函數(shù)進(jìn)行逼近[8],得到系統(tǒng)半離散化模型,最后對(duì)所得結(jié)果用Matlab進(jìn)行數(shù)值模擬,并得出相應(yīng)的模擬圖形,從直觀上驗(yàn)證理論研究結(jié)果的正確性.
具有兩種修復(fù)方法的復(fù)雜可修復(fù)系統(tǒng)(系統(tǒng)Ⅰ)的模型見(jiàn)圖1.
圖1 具有兩種修復(fù)方法的復(fù)雜可修復(fù)系統(tǒng)模型Fig.1 Repairable system model with two typesnof repair facilities
該模型可用積分-微分方程描述為
(1)
(2)
(3)
j=3,4
(4)
pj(0,t)=λjp2(t);j=3,4
(5)
p(0)=1,p1(0)=p2(0)=0,pj(x,0)=0;
j=3,4
(6)
記m0=λ1+λc,m1=λ2+μd,m2=λ3+λ4.其中pj(t)表示t時(shí)刻系統(tǒng)處于j狀態(tài)的概率,j=0為正常狀態(tài),j=1為退化狀態(tài),j=2為崩潰狀態(tài),j=3為大修狀態(tài),j=4為小修狀態(tài).pj(x,t)表示系統(tǒng)處于狀態(tài)j且已修時(shí)間為x的概率,j=3,4.λj是系統(tǒng)定常故障率,j=1為從正常狀態(tài)到退化狀態(tài),j=2為從退化狀態(tài)到崩潰狀態(tài),j=c為從正常狀態(tài)到崩潰狀態(tài).μd是系統(tǒng)在退化狀態(tài)時(shí)的定常修復(fù)率.μj(x)表示系統(tǒng)處于狀態(tài)j修復(fù)時(shí)間為x時(shí)的修復(fù)率,j=3,4,且滿足
在Banach空間中用抽象Cauchy問(wèn)題來(lái)描述這個(gè)系統(tǒng)狀態(tài)空間:
顯然X是Banach空間.定義算子A及其定義域:
則方程(1)~(6)可以描述成Banach空間X中一個(gè)抽象的Cauchy問(wèn)題:
(7)
p(0)=(1 0 0 0 0)T
(8)
下面構(gòu)造階梯函數(shù):
(9)
pn(0)=(1 0 0 0 0)T
(10)
由文獻(xiàn)[6]和[9]可知:A生成一個(gè)C0壓縮半群,再由生成C0半群的唯一性知此壓縮C0半群就是T(t).
首先估計(jì)線性算子A的預(yù)解式R(v;A)和線性算子An的預(yù)解式R(v;An),然后用Trotter定理來(lái)證明系統(tǒng)動(dòng)態(tài)解的逼近.
考慮方程(vI-A)p(x)=y(x),即
(11)
-λ1p0+(v+λ2+μd)p1=y1
(12)
-λcp0-λ2p1+(v+λ3+λ4)p2=y2
(13)
(14)
由方程(14)可得
(15)
令
由邊界條件,則有關(guān)于p0、p1、p2方程組如下:
(v+λ1+λc)p0-μdp1-(λ3σ3+λ4σ4)p2=y0+φ(y3)+φ(y4)
(16)
-λ1p0+(v+λ2+μd)p1=y1
(17)
-λcp0-λ2p1+(v+λ3+λ4)p2=y2
(18)
考慮關(guān)于p0、p1、p2方程組的系數(shù)矩陣D:
當(dāng)v>0時(shí)detD≠0,方程組(16)~(18)有唯一解[6,10],那么方程組(11)~(14)有唯一解,從而有R(vI-A)X=3.所以(vI-A)是閉算子,(vI-A)-1存在且有界[11].
由Gramer法則可得
其中d14=d11w3,d15=d11w4,d24=d21w3,d25=d21w4,d34=d31w3,d35=d31w4.
其中
k11=λ3d31w3, k12=λ3d32w3,
k13=λ3d33w3, k14=k15=λ3d31,
k21=λ4d31w4, k22=λ4d32w4,
k23=λ4d33w4, k24=k25=λ4d31
取
H=d11d12d13d14d15d21d22d23d24d25d31d32d33d34d35k11k12k13G k15φ4(τ)k21k22k23k24φ3(τ)G^?è?????????÷÷÷÷÷÷÷
其中
=k25φ4(τ)+G(y4)
因此A的預(yù)解式為
現(xiàn)在來(lái)證明系統(tǒng)修復(fù)率的逼近,只要證明R(v,An)y→R(v,A)y.
即證明
σnj→σj(n→∞)
考慮
則有
即
即
R(v;An)y→R(v;A)y(n→∞)
這樣就證明了系統(tǒng)動(dòng)態(tài)解的逼近.
下面利用數(shù)值計(jì)算方法,對(duì)上述結(jié)果進(jìn)行數(shù)值模擬,以期待驗(yàn)證理論結(jié)果的正確性,并以此說(shuō)明上述離散化方法的合理性.
為此假設(shè)故障率和修復(fù)率為常數(shù),即
λj=λc=λ(j=1,2,3,4),μd=μ3(x)=μ4(x)=μ.
并令
則系統(tǒng)(Ⅰ)轉(zhuǎn)化為一個(gè)常微分方程組(Ⅱ):
(19)
(20)
(21)
(22)
(23)
p0(0)=1,pj(0)=0;j=1,2,3,4
(24)
記m0=λ1+λc,m1=λ2+μd,m2=λ3+λ4.
下面用Matlab數(shù)學(xué)軟件求常微分方程組的數(shù)值解,此時(shí)令λ=0.5,μ=0.5,其結(jié)果如圖2所示.
(a) p0
(b) p1
(c) p2
(d) p3
(e) p4
圖2 系統(tǒng)Ⅰ的數(shù)值解(μ3(x)=μ4(x)=常數(shù))
Fig.2 Numerical solution of System Ⅰ (μ3(x)=μ4(x)=const)
由以上模擬圖形可以看出系統(tǒng)動(dòng)態(tài)解是存在的.這與以上證得的結(jié)論是相符的,從而也說(shuō)明了半離散化方法應(yīng)用于該模型是合理的.
本文通過(guò)半離散逼近算法將具有兩種修復(fù)方法的復(fù)雜可修復(fù)系統(tǒng)模型進(jìn)行合理離散并運(yùn)用Trotter逼近定理加以證明.同時(shí)在假設(shè)故障率和修復(fù)率為常數(shù)的前提下利用數(shù)值計(jì)算的方法對(duì)該模型進(jìn)行數(shù)值模擬,得到了該系統(tǒng)的數(shù)值解,并給出了相應(yīng)的數(shù)值模擬圖,從而更有效地利用計(jì)算機(jī)尋求數(shù)學(xué)問(wèn)題近似解,更好地解決數(shù)學(xué)問(wèn)題.
[1]GUPUR G. Asymptotic stability of the time-dependent solution of a reliability system [J]. Acta Analysis Functionalis Applicata, 2005, 7(4):219-316.
[2]LI Wei, CAO Jinhua. Some performance measures of transfer line consisting of two unreliable machines with reprocess rule [J]. Journal of Systems Science and Systems Engineering, 1998, 7(3):283-292.
[3]CHUNG Who-kee. A reliability analysis of ak-out-of-N:Gredundant system with the presence of chance common-cause shock failures [J]. Microelectronics Reliability, 1992, 32(10):1395-1399.
[4]DHILLON B S. Availability analysis of systems with two types of repair facilities [J]. Microelectronics Reliability, 1980, 20(5):679-686.
[5]張玉峰,喬 興. 在常規(guī)故障和臨界人為錯(cuò)誤條件下具有易損壞儲(chǔ)備部件復(fù)雜系統(tǒng)的可靠性分析[J]. 數(shù)學(xué)的實(shí)踐與認(rèn)識(shí), 2005, 35(7):195-206.
ZHANG Yufeng, QIAO Xing. Reliability analysis of a complex system with a deteriorating standby unit under common-cause failure and critical human error [J]. Mathematics in Practice and Theory, 2005, 35(7):195-206. (in Chinese)
[6]趙玉榮,喬 興,金雪梅. 有兩種修復(fù)方法的復(fù)雜可修復(fù)系統(tǒng)解的漸近穩(wěn)定性[J]. 數(shù)學(xué)的實(shí)踐與認(rèn)識(shí), 2008, 38(22):154-163.
ZHAO Yurong, QIAO Xing, JIN Xuemei. The asymptotic stability of the solution of a complex repairable system with two types of repair facilities [J]. Mathematics in Practice and Theory, 2008, 38(22):154-163. (in Chinese)
[7]周 莉,王偉華,張 敬. 兩相同部件冷貯備可修系統(tǒng)半離散化的研究[J]. 數(shù)學(xué)的實(shí)踐與認(rèn)識(shí), 2012, 42(5):127-132.
ZHOU Li, WANG Weihua, ZHANG Jing. The study on semi-dispersed algorithm of asymptotic stability of a cold standby repairable system [J]. Mathematics in Practice and Theory, 2012, 42(5):127-132. (in Chinese)
[8]陶有德,路振國(guó),范琳琳,等. 一類可修復(fù)計(jì)算機(jī)系統(tǒng)的數(shù)值計(jì)算[J]. 信陽(yáng)師范學(xué)院學(xué)報(bào)(自然科學(xué)版), 2013, 26(4):493-495.
TAO Youde, LU Zhenguo, FAN Linlin,etal. The numerical calculation of a repairable computer system [J]. Journal of Xinyang Normal University (Natural Science Edition), 2013, 26(4):493-495. (in Chinese)
[9]DAZY A. 線性算子半群及對(duì)偏微分方程的應(yīng)用[M]. 黃發(fā)倫,鄭 權(quán),譯. 成都:四川大學(xué)出版社, 1988.
DAZY A. Semigroups of Linear Operators and Their Applications to Partial Differential Equations [M]. HUANG Falun, ZHENG Quan, trans. Chengdu:Sichuan University Press, 1988. (in Chinese)
[10]徐厚寶,郭衛(wèi)華,于景元,等. 一類串聯(lián)可修復(fù)系統(tǒng)的穩(wěn)態(tài)解[J]. 應(yīng)用數(shù)學(xué)學(xué)報(bào), 2006, 29(1):46-52.
XU Houbao, GUO Weihua, YU Jingyuan,etal. The asymptotic stability of a series repairable system [J]. Acta Mathematicae Applicatae Sinica, 2006, 29(1):46-52. (in Chinese)
[11]張恭慶,郭懋正. 泛函分析講義[M]. 北京:北京大學(xué)出版社, 1990.
ZHANG Gongqing, GUO Maozheng. Functional Analysis [M]. Beijing:Beijing University Press, 1990. (in Chinese)
Study of solution of repairable system with two types of repair facilities
ZHOU Li*,LU Xuejuan,WANG Weihua
(College of Science, Qiqihar University, Qiqihar 161006, China )
The semi-discrete algorithm is applied to the repairable system with two types of repair facilities, the repairable rate in [0,x0] is discretized and the semi-discrete model of the system is acquired. Furthermore, by using the operator semi-group theory in functional analysis, the semi-discrete partial differential equation is transformed to the abstract Cauchy problems, i. e. the matrix ordinary differential equations. Then the solution of the matrix ordinary differential equations is proved to converge to the solution of the original equation according to Trotter approximate theorem. At last, because the failure rate and the repairable rate are constants, using Matlab the stability and the reliability of the system are proved and the numerical solution of the system is acquired, the corresponding graph trend is given. The results show that semi-discrete study of the repairable system model with two types of repair facilities can not only lay a theoretical foundation for the use of the computer for further numerical calculation, but also have practical value to analyze and study the reliability of the system.
repairable system; semi-discretization; convergence; numerical calculation
1000-8608(2017)04-0424-06
2017-03-09;
2017-06-05.
國(guó)家科技支撐計(jì)劃課題資助項(xiàng)目(2013BAK12B0803);黑龍江省教育廳基本業(yè)務(wù)專項(xiàng)理工面上項(xiàng)目(135109229).
周 莉*(1976-),女,碩士,副教授,E-mail:13796881349@139.com;蘆雪娟(1979-),女,博士,講師,E-mail:lujuan02@163.com;王偉華(1978-),女,碩士,副教授,E-mail:wangweihua8500@163.com.
TP391.9
A
10.7511/dllgxb201704014