南航双语矩阵论matrix theory第7章部分习题参考答案
第七章部分习题参考答案
Exercise 1
Show that a normal matrix A is Hermitian if its eigenvalues are all real.
Proof If A is a normal matrix, then there is a unitary matrix that diagonalizes A . That is, there is a unitary matrix U such that
H A UDU =
where D is a diagonal matrix and the diagonal elements of D are eigenvalues of A . If eigenvalues of A are all real, then
()H H H H H H A UDU UD U UDU A ====
Therefore, A is Hermitian.
Exercise 2
Let A and B be Hermitian matrices of the same order. Show that AB is Hermitian if and only if AB BA =. Proof
If AB BA =, then ()()H H H H AB BA A B AB ===. Hence, AB is Hermitian. Conversely, if AB is Hermitian, then ()H AB AB =. Therefore, H H AB B A BA ==.
Exercise 3
Let A and B be Hermitian matrices of the same order. Show that A and B are similar if they have the same characteristic polynomial.
Proof Since matrix A and B have the same characteristic polynomial, they have the same eigenvalues 12,,,n λλλ . There exist unitary matrices U and V such that
12diag(,,,)H n U AU λλλ= , 12diag(,,,)H n V BV μμμ= .
Thus,
H H U AU V BV =. (11,H H U U V V --==)
That is 1()H H UV AUV B -=. Hence, A and B are similar.
Exercise 4
Let A be a skew-Hermitian matrix, i.e., H A A =-, show that (a) I A - and I A + are invertible.
(b) 1()()I A I A --+ is a unitary matrix with eigenvalues not equal to 1-. Proof of Part (a)
Method 1: (a) since H A A =-, it follows that
()()H I A I A I AA I A A -+=-=+
For any x 0≠
()()0x x x x x x x x x x H H H H H H H I A A A A A A +=+=+>
Hence, ()()I A I A -+ is positive definite. It follows that ()()I A I A -+ is invertible. Hence, both I A - and I A + are invertible.
Method 2:
If I A - is singular, then there exists a nonzero vector x such that
()x 0I A -=. Thus, x x A =,
x x x x H H A =. (1)
Since x x H is real, it follows that
()x x x x H H H A =.
That is x x x x H H H A =. Since H A A =-, it follows that
x x x x H H A -= (2)
Equation (1) and (2) implies that 0x x H =. This contradicts the assumption that x is nonzero. Therefore, I A - is invertible.
Method 3:
Let λ be an eigenvalue of A and x be an associated eigenvector. x x A λ=
x x x x H H A λ=. ()x x x x x x x x x x x x
H H H H H H H H A A A λλ===-=-
Hence, λ is either zero or pure imaginary. 1 and 1- can not be eigenvalues of A . Hence, I A -
and I A + are invertible.
Method 4: Since H A A =-, A is normal. There exists a unitary matrix U such that 12diag(,,,)H n U AU λλλ=
12()()diag(,,,)H H H H H H n U AU U A U U AU ==-= 12diag(,,,)n λλλ= 12diag(,,,)n λλλ- Each j λ is pure imaginary or zero.
12(diag(,,,))H n I A U I U λλλ-=-
12diag(1,1,,1))H n I A U U λλλ-=---
Since 10i λ-≠ for 1,2,,j n = , det ()0I A -≠. Hence, I A - is invertible. Similarly, we can prove that I A + is invertible.
Proof of Part (b) Method 1:
Since ()()()()I A I A I A I A +-=-+, it follows that
11[()()]()()H I A I A I A I A ---+-+
11()()()()H H I A I A I A I A --=+--+ ( Note that 11()()H H P P --= if P is nonsingular.)
11()()()()I A I A I A I A --=-+-+ 11()()()()I A I A I A I A I --=--++=
Hence, 1()()I A I A --+ is a unitary matrix. Denote 1()()B I A I A -=-+.
Since 111(1)(1)()()()()2()I B I I A I A I A I A I A I A -----=---+=-++-+=-+,
1det()(2)det[()]0n I B I A ---=-+≠
Hence, 1- can not be an eigenvalue of 1()()I A I A --+. Method 2:
By method 4 of the Proof of Part (a),
12diag(1,1,,1))H n I A U U λλλ-=---
12diag(1,1,,1))H n I A U U λλλ+=+++
1()()I A I A --+12
12111diag(
,,,))111H n n
U U λλλλλλ---=+++ The eigenvalues of 1()()I A I A --+ are
12
12111,,,111n n
λλλλλλ---+++ , which are all not equal to 1-.
Method 3: Since ()()()()I A I A I A I A +-=-+, it follows that
11()()()()I A I A I A I A ---+=+-
If 1- is an eigenvalue of 1()()I A I A --+, then there is a nonzero vector x , such that
1()()x x I A I A --+=-. That is 1()()x x I A I A -+-=-.
It follows that
()()x x I A I A -=-+.
This implies that x 0=. This contradiction shows that 1- can not be an eigenvalue of
1()()I A I A --+.
Exercise 6
If H is Hermitian, show that i I H - is invertible, and 1(i )(i )U I H I H -=+- is unitary. Proof Let i A H =-. Then A is skew-Hermitian. By Exercises #4, I A - and I A + are invertible, and 1()()U I A I A -=-+ is unitary. This finishes the proof.
Exercise 7
Find the Hermitian matrix for each of the following quadratic forms. And reduce each quadratic form to its canonical form by a unitary transformation (a) 12312131213(,,)i i f x x x x x x x x x x x =+-+ Solution
(
)
11231
2
3230i 1(,,)i 00100x f x x x x x x x x ???? ???=- ??? ???????, 0i 1i 00100A ?? ?=- ? ???
3d e t ()2I A λλλ-=-. Eigenvalues of A
are 1λ=
2λ=30λ=.
Associated unit eigenvectors are
1i 1,)22u T =-
, 2i 1
,)22
u T =-, and
3u T =, respectively. 123,,u u u form an orthonormal set.
Let 123(,,)u u u U =, and x y U =. Then we obtain the canonical form
1122y y
Exercise 9
Let A and B be Hermitian matrices of order n , and A be positive definite. Show that AB is
similar to a real diagonal matrix.
Proof Since A is positive definite, there exists an nonsingular Hermitian matrix P such that H A PP = 1()H H AB PP B P P BP P -==
AB is similar to H P BP . Since H P BP is Hermitian, it is similar to a real diagonal matrix. Hence, AB is similar to a real diagonal matrix.
Exercise 10
Let A be an Hermitian matrix of order n . Show that there exists a real number 0t such that t I A +is positive definite.
Proof 1: The matrix t I A + is Hermitian for real values of t . If the eigenvalues of A are
12,n λλλ ,,
, then the eigenvalues of t I A +are 12,,n t t t λλλ+++ ,. Let 12max{,,}n t λλλ> ,
Then the eigenvalues of t I A + are all positive. And hence, tI A +is positive definite.
Proof 2: The matrix t I A + is Hermitian for real values of t . Let r A be the leading principle minor of A of order r .
d e t (
)r r r I A t +=+terms involving lower powers in t . Hence, det()r r t I A + is positive for sufficiently large t .
Thus, if t is sufficiently large, all leading principal minors of t I A + will be positive.
That is, there exists a real number 0t such that det()r r t I A + is positive for 0t t > and for each r . Thus t I A + is positive definite for 0t t >.
Exercise 11 Let
11
121222H A A A A A ??
= ???
be an Hermitian positive definite matrix. Show that 1122det()det()det()A A A ≤
Proof We first prove that if A is Hermitian positive definite and B is Hermitian semi-positive
definite, then det()det()A B A +≥. Since A is positive definite, there exists a nonsingular hermitian matrix P such that
H
A P P =
11(())H H A B P I P B P P --+=+ 11det()det()det(())H A B A I P B P --+=+
11()H I P B P --+ is positive semi- definite. Its eigenvalues are all greater than or equal to 1.
Thus
11det(())1H I P B P --+≥
1
11121112112111222H H I O A A I A A A A I A A O I --??
-?
???
? ???-??????
1
11121111
121
1
2212111222121112H H A A A O I A A O A A A A O A A A A O I ---??-????
== ? ? ?--????
?? 122121112H A A A A -- is positive definite, and 1
121112H A A A - is positive semi-definite, and
1
112212
1112det()det()det()H A A A A A A -=- Hence, 111
222212111212111222121112det()det()det(H H H A A A A A A A A A A A A ---=-+≥-)
This finishes the proof.
Exercise 12
Let A be a positive definite Hermitian matrix of order n . Show that the element in A with the largest norm must be in the main diagonal.
Proof Let ()ij A a =. Suppose that 00
i j a is of the largest norm, where 00i j ≠. Consider the
principal minor 00
000000i i i j i j j j a a a a ??
? ???
. It must be positive definite since A is positive definite. (Recall that an Hermitian matrix is positive definite iff all its principal minors are positive.) Thus, 00
0000
00det 0i i i j i j j j a a a a ??
?> ???
. On the other hand, 00
00000000
00
002
det 0i i i j i i j j i j i j j j a a a a a a a ??
?=-≤ ???
since 00i j a is of the largest norm.
(Remark: The diagonal elements in an Hermitian matrix must be real.)
This contradiction implies that the element in A with the largest norm must be in the main diagonal.
南航矩阵论等价关系
Student’s Name: Student’s ID No.: College Name: The study of Equivalence Relations Abstract According to some relative definitions and properties, to proof that if B can be obtained from A by performing elementary row operations on A, ~ is an equivalence relation, and to find the properties that are shared by all the elements in the same equivalence class. To proof that if B is can be obtained from A by performing elementary operations, Matrix S A ∈ is said to be equivalent to matrix S B ∈, and ~A B means that matrix S A ∈ is similar to S B ∈, if let S be the set of m m ? real matrices. Introduction The equivalence relations are used in the matrix theory in a very wide field. An equivalence relation on a set S divides S into equivalence classes. Equivalence classes are pair-wise disjoint subsets of S . a ~ b if and only if a and b are in the same equivalence class.This paper will introduce some definitions and properties of equivalence relations and proof some discussions. Main Results Answers of Q1 (a) The process of the proof is as following,obviously IA=A,therefore ~ is reflexive;we know B can be obtained from A by performing elementary row operations on A,we assume P is a matrix which denote a series of elementary row operations on A.Then ,we have PA=B,(A~B),and P is inverse,obviously we have A=P -1B,(B~A).So ~ is symmetric.We have another matrix Q which denote a series of elementary row operations on B,and the result is C,so we have QB=C.And we can obtain QB=Q(PA)=QPA=C,so A~C.Therefore,~ is transitive. Hence, ~ is an equivalence relation on S . (b) The properties that are shared by all the elements in the same equivalence class are as followings: firstly,the rank is the same;secondly,the relation of column is not changed;thirdly,two random matrices are row equivalent;fourthly,all of the matrices
2016矩阵论试题
第 1 页 共 6 页 (A 卷) 学院 系 专业班级 姓名 学号 (密封线外不要写姓名、学号、班级、密封线内不准答题,违者按零分计) …………………………………………密…………………………封……………………………………线………………………………… 考试方式:闭卷 太原理工大学 矩阵分析 试卷(A ) 适用专业:2016级硕士研究生 考试日期:2017.1.09 时间:120 分钟 共 8页 一、填空选择题(每小题3分,共30分) 1-5题为填空题: 1. 已知??? ? ? ??--=304021101A ,则1||||A =。 2. 设线性变换1T ,2T 在基n ααα ,,21下的矩阵分别为A ,B ,则线性变换212T T +在基n ααα ,,21下的矩阵为_____________. 3.在3R 中,基T )2,1,3(1--=α,T )1,1,1(2-=α,T )1,3,2(3-=α到基T )1,1,1(1=β, T )3,2,1(2=β,T )1,0,2(3=β的过度矩阵为A = 4. 设矩阵??? ? ? ??--=304021101A ,则 5432333A A A A A -++-= . 5.??? ? ? ? ?-=λλλλλ0010 01)(2A 的Smith 标准形为 6-10题为单项选择题: 6.设A 是正规矩阵,则下列说法不正确的是 ( ). (A) A 一定可以对角化; (B )?=H A A A 的特征值全为实数; (C) 若E AA H =,则 1=A ; (D )?-=H A A A 的特征值全为零或纯虚数。 7.设矩阵A 的谱半径1)( 南京航空航天大学2012级硕士研究生 二、(20分)设三阶矩阵,,. ????? ??--=201034011A ????? ??=300130013B ???? ? ??=3003003a a C (1) 求的行列式因子、不变因子、初等因子及Jordan 标准形; A (2) 利用矩阵的知识,判断矩阵和是否相似,并说明理由. λB C 解答: (1)的行列式因子为;…(3分)A 2121)1)(2()(,1)()(--===λλλλλD D D 不变因子为; …………………(3分)2121)1)(2()(,1)()(--===λλλλλd d d 初等因子为;……………………(2分) 2)1(,2--λλJordan 标准形为. ……………………(2分) 200011001J ?? ?= ? ??? (2) 不相似,理由是2阶行列式因子不同; …………………(5分) 0,a = 相似,理由是各阶行列式因子相同. …………………(5分) 0,a ≠共 6 页 第 4 页 三、(20分)已知线性方程组不相容. ?? ???=+=+++=++1,12,1434321421x x x x x x x x x (1) 求系数矩阵的满秩分解; A (2) 求广义逆矩阵; +A (3) 求该线性方程组的极小最小二乘解. 解答:(1) 矩阵,的满秩分解为 ???? ? ??=110021111011A A . …………………(5分)10110111001101A ??????=?????????? (2) . ……………………(10分)51-451-41-52715033A +?? ? ?= ? ??? (3) 方程组的极小最小二乘解为. …………(5分)2214156x ?? ? ?= ? ??? 共 6 页 第 5 页 Solution Key to Some Exercises in Chapter 3 #5. Determine the kernel and range of each of the following linear transformations on 2P (a) (())'()p x xp x σ= (b) (())()'()p x p x p x σ=- (c) (())(0)(1)p x p x p σ=+ Solution (a) Let ()p x ax b =+. (())p x ax σ=. (())0p x σ= if and only if 0ax = if and only if 0a =. Thus, ker(){|}b b R σ=∈ The range of σis 2()P σ={|}ax a R ∈ (b) Let ()p x ax b =+. (())p x ax b a σ=+-. (())0p x σ= if and only if 0ax b a +-= if and only if 0a =and 0b =. Thus, ker(){0}σ= The range of σis 2()P σ=2{|,}P ax b a a b R +-∈= (c) Let ()p x ax b =+. (())p x bx a b σ=++. (())0p x σ= if and only if 0bx a b ++= if and only if 0a =and 0b =. Thus, ker(){0}σ= The range of σis 2()P σ=2{|,}P bx a b a b R ++∈= 备注: 映射的核以及映射的像都是集合,应该以集合的记号来表达或者用文字来叙述. #7. Let be the linear mapping that maps 2P into 2R defined by 10()(())(0)p x dx p x p σ?? ?= ??? ? Find a matrix A such that ()x A ασαββ??+= ??? . Solution 1(1)1σ??= ??? 1/2()0x σ?? = ??? 11/211/2()101 0x ασαβαββ????????+=+= ? ? ??????????? Hence, 11/210A ??= ??? #10. Let σ be the transformation on 3P defined by (())'()"()p x xp x p x σ=+ a) Find the matrix A representing σ with respect to 2[1,,]x x b) Find the matrix B representing σ with respect to 2[1,,1]x x + c) Find the matrix S such that 1B S AS -= d) If 2012()(1)p x a a x a x =+++, calculate (())n p x σ. Solution (a) (1)0σ= 1) 一组基为q = .维数为3. 3) 南京航空航天大学双语矩阵论期中考试参考答案(有些答案可能有问题) Q1 1解矩阵A 的特征多项式为 A-2 3 -4 4I-A| =-4 2+6 -8 =A 2(/l-4) -6 7 A-8 所以矩阵A 的特征值为4 =0(二重)和/^=4. 人?2 3 由于(4-2,3)=1,所以D| (人)二1.又 彳 人+6=“2+4人=?(人) 4-2 3 、=7人+4=代(人)故(们3),代3))=1 ?其余的二阶子式(还有7个)都包含因子4, -6 7 所以 D? 3)=1 .最后 det (A (/L))=42(人.4),所以 D 3(A)=/l 2 (2-4). 因此矩阵A 的不变因子为d, (2) = d 2(2) = l, d 3 (2) = r (2-4). 矩阵A 的初等因子为人2, 2-4. 2解矩阵B 与矩阵C 是相似的.矩阵B 和矩阵C 的行列式因子相同且分别为9 3)=1 , D 2(/i)=A 2-/l-2 .根据定理:两矩阵相似的充分必要条件是他们有相同的行列式因子. 所以矩阵B 与矩阵c 相似. Q2 2)设k 是数域p 中任意数,a, 0, /是v 中任意元素.明显满足下而四项. (") = (",a) ; (a+月,/) = (",/) + (”,刃;(ka,/3) = k(a,/3) ; (a,a)>0, 当且仅当Q = 0时(a,a) = ().所以(。,/?)是线性空间V 上的内积. 利 用Gram-Schmidt 正交化方法,可以依次求出 ,p 2 =%-(%'5)与= 层=%-(%,弟与一(%,弓)役=南航矩阵论2013研究生试卷及答案
南航双语矩阵论 matrix theory第三章部分题解
南航矩阵论期中考试参考答案.doc
2016矩阵论试题A20170109 (1)