闭复区间矩阵及其应用初论

2014-09-10 17:44赵志青石春
电脑知识与技术 2014年6期

赵志青 石春

摘要:区间矩阵在工程领域,图像识别和力学中的应用和范围越来越受到人们的关注,其在处理多维数据,尤其在不确定方程的求解中发挥了重要的作用,该文在在区间矩阵的基础之上,给出了闭复区间矩阵,研究了其性质,最后将其应用到复区间数值的方程组求解中,为其更深入的研究奠定了扎实的基础,丰富和发展了矩阵理论及相关学科。

关键词:区间数;复区间值;区间矩阵;闭复区间矩阵;复区间值方程组

中图分类号:TP18 文献标识码:A 文章编号:1009-3044(2014)06-1309-05

1 预备知识

区间数

定义1[1] 设[R=-∞,+∞]为实数空间,称R上的有限区间[X=X-,X+]为区间数,区间数记作[IR]

定义2[1] 任意的[X,Y∈IR],有:

[X∨Y=X-∨Y-,X+∨Y+][X∧Y=X-∧Y-,X+∧Y+]

[X+Y=X-+Y-,X++Y+][X-Y=X--Y+,X+-Y-]

[X?Y=X-Y-∧X-Y+∧X+Y-∧X+Y+,X-Y-∨X-Y+∨X+Y-∨X+Y+]

[ab=a-b-∧a-b+∧a+b-∧a+b+,a-b-∨a-b+∨a+b-∨a+b+,0?b]

[kX=kX-,kX+0kX+,kX-k?0k=0k?0] [1b=Δ1b-∧1b+,1b-∨1b+]

定义3 设[a=a-,a+,b=b-,b+∈IR].

1)如果[a-≤b-,a+≤b+],称a小于或等于b.若[a=a-,a+=0-,0+],则称b大余或等于a.

1.2 闭区间复数

定义4[1] 设C为复数域,对任意闭区间数

[X=X-,X+,Y=Y-,Y+][∈IR]

称复有界闭集,

[Z=X+iY=x+iy∈C|x∈X,y∈Y]

为闭复区间数,其中[i=-1],用[IC]表示[C]上闭复区间数全体,即[IC=z=x+iy|x,y∈IR].

定义5[1]设*是复数域C上的二元运算,对于任意的

[Zk=Xk+iYk=X-k,X+k+iY-k,Y+k∈Ic][k=1,2Ic]

上的扩展运算定义为

[Z1*Z2=ΔZ|?Z1,Z2∈Z1×Z2,Z=Z1*Z2]

定义6[1]共轭闭复区间数

称[Z*=X-iY=x-iy|x∈X,y∈Y]为[Z]的共轭闭复区间数。

称[Z=X2+Y212=x2+y212|x∈X,y∈Y]为Z上的模。

定义7[1] 对任意的[Zk=Xk+iYk=X-k,X+k+iY-k,Y+k∈ICk=1,2]

有如下运算:

[Z1+Z2=X-1,X+1+iY-1,Y+1+X-2,X+2+iY-2,Y+2][=X-1+X-2,X+1+X+2+iY-1+Y-2,Y+1+Y+2]

[Z1-Z2=X-1,X+1+iY-1,Y+1-X-2,X+2+iY-2,Y+2][=X-1-X+2,X+1-X-2+iY-1-Y+2,Y+1-Y-2]

[Z1Z2=X1+iY1X2+iY2][=X1X2+iX1Y2+iX2Y1-Y1Y2][=X1X2-Y1Y2+iX1Y2+X2Y1]

(其中,[Xi?Yj=Xi-Yj-∧Xi-Yj+∧Xi+Yj-∧Xi+Yj+,Xi-Yj-∨Xi-Yj+∨Xi+Yj-∨Xi+Yj+,i∈1,2,j∈1,2])

[kZ=kx+yi=kx+kyi=kx-,kx++ky-,ky+i0kx+,kx-+ky+,ky-ik>0k=0k<0]

1.3 复矩阵

复矩阵指的是元素中含有复数的矩阵.其中不管这个矩阵中含有多少个复数,只要这个矩阵中含有复数,那么这个矩阵就是复矩阵.如果这个矩阵中不含有复数,那么这个矩阵就不是复矩阵.

2 闭复区间矩阵

2.1 闭复区间矩阵的定义

定义1 如果[m×n]个元素均为闭复区间数,[Zij∈IC][i=1,2,..........m,j=1,2,.............n]则由[m×n]个元素构成的[m×n]阶矩阵

[E=Z11Z12......Z1nZ21Z22......Z2n..........Zm1Zm2......Zmn]

记为[EC],表示C上的所有闭复区间矩阵.

特别的当[m=n]时,则称为[n]阶闭复区间矩阵记为[Enc]

定义2 形如

[1+i00...001+i0...0...............00...1+i000...01+i]

的矩阵称为单位矩阵.这里的[1+i=1-,1++1-,1+i]

形如

[00...0000...00...............00...0000000]

的矩阵称为零矩阵.这里的[0=0-,0++0-,0+i]

2.2 闭复矩阵的运算

定义3 对任意的C上的闭复区间矩阵

[E=Z11Z12......Z1nZ21Z22......Z2n..........Zm1Zm2......Zmn]

[F=Y11Y12......Y1nY21Y22......Y2n..........Ym1Ym2......Ymn]

则有

[1][E+F=][Z11Z12......Z1nZ21Z22......Z2n..........Zm1Zm2......Zmn][±Y11Y12......Y1nY21Y22......Y2n..........Ym1Ym2......Ymn][=Z11±Y11Z12±Y12......Z1n±Y1nZ21±Y21Z22±Y22......Z2n±Y2n..........Zm1±Ym1Zm2±Ym2......Zmn±Ymn]

特别的当E的列数等于F的行数时有

[EF=][Z11Z12......Z1nZ21Z22......Z2n..........Zm1Zm2......Zmnm×n][Y11Y12......Y1mY21Y22......Y2m..........Yn1Yn2......Ynmn×m]

[Z11Y11+Z12Y21+...+Z1nYn1Z11Y21+Z12Y22+...+Z1nYn2...Z11Y1n+Z12Y2n+...+Z1nYnm........Zm1Y11+Zm2Y21+...+ZmnYn1Zm1Y12+Zm2Y22+...+ZmnYn2...Zm1Y1n+Zm2Y2n+...+ZmnYnmm×m]

2)对于任意的[k∈IR]有

[kE=kZ11kZ12......kZ1nkZ21kZ22......kZ2n..........kZm1kZm2......kZmn]

2)闭复区间矩阵的转置

[Z11Z12......Z1nZ21Z22......Z2n..........Zm1Zm2......ZmnT=Z11Z21......Zn1Z12Z22......Zn2..........Z1nZ2m......Znm]

4) 闭复区间矩阵的共轭矩阵

[Z11Z12......Z1nZ21Z22......Z2n..........Zm1Zm2......Zmn为Z11Z12......Z1nZ21Z22......Z2n..........Zm1Zm2......Zmn]

定理1 (基本性质)

设[E1,E2,E3∈Ec],[k,l∈IR],有

1)(加的交换律)[E1+E2=E2+E1]

2) (加的结合律) [E1+E2+E3=E1+E2+E3]

3) (加零)[E1+0=E1]

4)(加负)[E1+-E1=0]

5) (分配律)[k+lE1=kE1+lE2]

6) (分配律)[kE1+E2=kE1+kE2]

7) [Z=Z],[z1+z2=z1+z2]

8) [z1z2=z1z2],[z1z2=z1z2][z1≠0]

2.3 闭复区间矩阵的行列式

定义4 如果[m×n]个元素均为闭复区间数,[Zij∈IC][i,j=1,2,..........n]则由[n×n]个元素构成的n阶行列式

[Z11Z12......Z1nZ21Z22......Z2n..........Zn1Zn2......Znn]

是[IC]中唯一确定的数:

1)当[n=1]时,[Z11=Z11]

2)当[n>1]时,

[Z11Z12...Z1nZ21Z22...Z2n............Zn1Zn2...Znn=Z11Z22...Z2nZ32...Z3n.........Zn2...Znn-Z21Z12...Z1nZ32...Z3n.........Zn2...Znn+......+-1n-1Zn1Z12...Z1nZ22...Z2n.........Zn-12...Zn-1n]

2.4 闭复区间矩阵的克莱姆法则

定理4 如果线性方程组

[Z11*x1+Z12*x2+......+Z1n*xn=b1Z21*x1+Z22*x2+......+Z2n*xn=b2...Zn1*x1+Zn2*x2+......+Znn*xn=bn]

的系数行列式

[d=Z11Z12...Z1nZ21Z22...Z2n.........................................................Zn1Zn2...Znn≠0]

则此方程存在唯一解:

[x1=d1d,x2=d2d,......,xn=dnd]

其中

[dj=Z11Z12...Z1j-1b1Z1j+1...Z1nZ21Z22...Z2j-1b2Z2j+1...Z2n.........................................................Zn1Zn2...Znj-1bnZnj+1...Znn,j=1,2,...,n.]

例1 解下列闭复区间方程组。

[13+25ix1+24+35ix2=57+46i27+36ix1+79+68ix2=810+1112i]

解:

[d=13+25i24+35i27+36i79+68i=13+25i79+68i-24+35i27+36i=1379+1368i+2579i-2568-2427+2436i+3527i-3536]

[=1×7∧1×9∧3×7∧3×9,1×7∨1×9∨3×7∨3×9+1×6∧1×8∧3×6∧3×8,1×6∨1×8∨3×6∨3×8i+2×7∧2×9∧5×7∧5×9,2×7∨2×9∨5×7∨5×9i-2×6∧2×8∧5×6∧5×8,2×6∨2×8∨5×6∨5×8-{2×2∧2×7∧4×2∧4×7,2×2∨2×7∨4×2∨4×7+2×3∧2×6∧4×3∧4×6,2×3∨2×6∨4×3∨4×6i+2×3∧2×6∧4×3∧4×6,2×3∨2×6∨4×3∨4×6i-3×3∧3×6∧5×3∧5×6,3×3∨3×6∨5×3∨5×6}]

[=(7∧9∧21∧27,7∨9∨21∨27+6∧8∧18∧24,6∨8∨18∨24i+14∧18∧35∧45,14∨18∨35∨45i-12∧16∧30∧40,12∨16∨30∨40)-(4∧14∧8∧28,4∨14∨8∨28+6∧12∧12∧24,6∨12∨12∨24i+6∧21∧10∧35,6∨21∨10∨35i-9∧18∧15∧30,9∨18∨15∨30)=727+624i+1445i-1240-428+624i+635i-930]

[=727-1240+624-1445i-428-930+624-635i=7-4027-12+2069i-4-3028-9+6-3524-6i=-3315+2069i--2619+-2918i=-3315--2619+2069--2918i=-33-1915+26+20-1869+29i=-5241+298i]

[d≠0]

[d1=57+46i24+35i810+1112i79+68i]

[=57+46i79+68i-24+35i810+1112i]

[=5779+5768i+4679i-4668-24810+241112i+35810i-351112=(5×7∧5×9∧7×7∧7×9,5×7∨5×9∨7×7∨7×9+5×6∧5×8∧7×6∧7×8,5×6∨5×8∨7×6∨7×8i+4×7∧4×9∧6×7∧6×9,4×7∨4×9∨6×7∨6×9i-4×6∧4×8∧6×6∧6×8,4×6∨4×8∨6×6∨6×8)-(2×8∧2×10∧4×8∧4×10,2×8∨2×10∨4×8∨4×10+2×11∧2×12∧4×11∧4×12,2×11∨2×12∨4×11∨4×12i+3×8∧3×10∧5×8∧5×10,3×8∨3×10∨5×8∨5×10i-3×11∧3×12∧5×11∧5×12,3×11∨3×12∨5×11∨5×12)=-20-5+104208i]

[d2=-5534+-5763i]

[x1=-11.471.984+-4.014.16i]

[x2=-4.243.685+-3.9253.1i]

参考文献:

[1] 马生全.模糊复分析理论基础[M].北京:科学出版社,2010.

[2] 樊恽,钱吉林,岑嘉评,等.代数学大词典[M].武汉:华中师范大学出版社,1994.

[3] 钟玉泉.复变函数论[M].3版.北京:高等教育出版社,2004

[4] KATARINA C.Eigenvectors of interval matrices over max-plus algebra [J].Discrete Applied Mathematics, 2005, 150 (1/2/3):2-15.

[5] Jiang, Chungli. Sufficient condition for the asymptotic stability of interval matric es [J]. INT. J. Control.,1987, 46(5):1803-1810.

[6] SoH Y C, Evans R J. Stability analysis of interval. Matrices-continuous and discrete systems[J]. International Journal of Control, 1988, 47(1):25-32.

[7] Man Sour M. Simpliyed sufficient Conditions fo the asymptotic Stability of interval Martices[J]. International Journal of Control, 1989, 50(1):443-444.

[8] Wang jingguo. Necessary and sufficient conditions for stability of a Matrix polytope with Normal Vertex Matrices [J]. Automatic, 1991, 27(5):887-888.

[9] 川陈景良,陈向晖.特殊矩阵[M].北京:清华大学出版社,2001.