Section 7.9

General Theory of Matrices

[36/83%]
Show
LaTeX
Problem 7.9.1

Prove that a commutative C\mathbb {C}-algebra of 2×22 \times 2 complex matrices has dimension at most 22 over C\mathbb {C}.

?
Problem 7.9.2

Prove the following three statements about real n×nn \times n matrices.

?
(1)

If AA is an orthogonal matrix whose eigenvalues are all different from −1-1, then I+AI+A is nonsingular and S=(I−A)(I+A)−1S = (I-A)(I+A)^{-1} is skew-symmetric.

(2)

If SS is a skew-symmetric matrix, then A=(I−S)(I+S)−1A = (I-S)(I+S)^{-1} is an orthogonal matrix with no eigenvalue equal to −1-1.

(3)

The correspondence A↔SA \leftrightarrow S from Parts 1 and 2 is one-to-one.

Problem 7.9.3

Let BB denote the matrix

(a000b000c) \begin{pmatrix} a & 0 & 0 \\ 0 & b & 0 \\ 0 & 0 & c \end{pmatrix}

where aa, bb, and cc are real and ∣a∣\left|a\right|, ∣b∣\left|b\right|, and ∣c∣\left|c\right| are distinct. Show that there are exactly four symmetric matrices of the form BQBQ, where QQ is a real orthogonal matrix of determinant 11.

?
Problem 7.9.4

Let PP be a n×nn \times n real matrix such that xtPy=−ytPxx^{t}Py = -y^{t}Px for all column vectors x,yx, y in Rn\mathbb {R}^{n}. Prove that PP is skew-symmetric.

?
Problem 7.9.5

Let AA be a real skew-symmetric matrix (Aij=−AjiA_{ij} = -A_{ji}). Prove that AA has even rank.

?
Problem 7.9.6

A real symmetric n×nn \times n matrix AA is called positive semi-definite if xtAx≥0x^{t}Ax \geq 0 for all x∈Rnx \in \mathbb {R}^{n}. Prove that AA is positive semi-definite if and only if tr⁡AB≥0\operatorname {tr} AB \geq 0 for every real symmetric positive semi-definite n×nn \times n matrix BB.

?
Problem 7.9.7

Suppose that AA and BB are real matrices such that

At=A,vtAv≥0 A^{t} = A, \qquad v^{t}Av \geq 0

for all v∈Rnv \in \mathbb {R}^{n} and

AB+BA=0. AB+BA = 0.

Show that AB=BA=0AB = BA = 0 and give an example where neither AA nor BB is zero.

?
Problem 7.9.8

Let AA and BB be n×nn \times n matrices over R\mathbb {R} such that A+BA+B is invertible. Prove that

A(A+B)−1B=B(A+B)−1A. A(A+B)^{-1}B = B(A+B)^{-1}A.
?
Problem 7.9.9

Suppose AA is a real n×nn \times n matrix.

?
(1)

Is it true that AA must commute with its transpose?

(2)

Suppose the columns of AA (considered as vectors) form an orthonormal set; is it true that the rows of AA must also form an orthonormal set?

Problem 7.9.10

Let M1=(3214)M_{1} = \begin{pmatrix} 3 & 2 \\ 1 & 4 \end{pmatrix}, M2=(57−3−4)M_{2} = \begin{pmatrix} 5 & 7 \\ -3 & -4 \end{pmatrix}, M3=(56.9−3−4)M_{3} = \begin{pmatrix} 5 & 6.9 \\ -3 & -4 \end{pmatrix}. For which (if any) ii, 1≤i≤31 \leq i \leq 3, is the sequence (Min)(M_{i}^{n}) bounded away from ∞\infty? For which ii is the sequence bounded away from 00?

?
Problem 7.9.11

Let AA be an n×nn \times n complex matrix, all of whose eigenvalues are equal to 11. Suppose that the set {An∣n=1,2,…}\left\{ A^{n} \mid n = 1, 2, \ldots \right\} is bounded. Show that AA is the identity matrix.

?
Problem 7.9.12

Consider the complex 3×33 \times 3 matrix

A=(a0a1a2a2a0a1a1a2a0), A = \begin{pmatrix} a_{0} & a_{1} & a_{2} \\ a_{2} & a_{0} & a_{1} \\ a_{1} & a_{2} & a_{0} \end{pmatrix},

where a0,a1,a2∈Ca_{0}, a_{1}, a_{2} \in \mathbb {C}.

?
(1)

Show that A=a0I3+a1E+a2E2A = a_{0}I_{3}+a_{1}E+a_{2}E^{2}, where

E=(010001100). E = \begin{pmatrix} 0 & 1 & 0 \\ 0 & 0 & 1 \\ 1 & 0 & 0 \end{pmatrix}.
(2)

Use Part 1 to find the complex eigenvalues of AA.

(3)

Generalize Parts 1 and 2 to n×nn \times n matrices.

Problem 7.9.13

Let AA be a n×nn \times n real matrix.

?
(1)

If the sum of each column element of AA is 11 prove that there is a nonzero column vector xx such that Ax=xAx = x.

(2)

Suppose that n=2n=2 and all entries in AA are positive. Prove there is a nonzero column vector yy and a number λ>0\lambda > 0 such that Ay=λyAy = \lambda y.

Problem 7.9.14

Let the real 2n×2n2n \times 2n matrix XX have the form

(ABCD) \begin{pmatrix} A & B \\ C & D \end{pmatrix}

where A,B,CA, B, C, and DD are n×nn \times n matrices that commute with one another. Prove that XX is invertible if and only if AD−BCAD-BC is invertible.

?
Problem 7.9.15

Let GL2(C)GL_{2}(\mathbb {C}) denote the group of invertible 2×22 \times 2 matrices with coefficients in the field of complex numbers. Let PGL2(C)PGL_{2}(\mathbb {C}) denote the quotient of GL2(C)GL_{2}(\mathbb {C}) by the normal subgroup {(λ00λ)∣λ∈C∗}\left\{ \begin{pmatrix} \lambda & 0 \\ 0 & \lambda \end{pmatrix} \mid \lambda \in \mathbb {C}^{*}\right\}. Let nn be a positive integer, and suppose that a,ba, b are elements of PGL2(C)PGL_{2}(\mathbb {C}) of order exactly nn. Prove that there exists c∈PGL2(C)c \in PGL_{2}(\mathbb {C}) such that cac−1cac^{-1} is a power of bb.

?
Problem 7.9.16

Let B=(bij)i,j=120B = (b_{ij})_{i,j=1}^{20} be a real 20×2020 \times 20 matrix such that

bii=0for 1≤i≤20, b_{ii} = 0 \quad \text{for } 1 \leq i \leq 20, bij∈{1,−1}for 1≤i,j≤20,  i≠j. b_{ij} \in \left\{ 1,-1\right\} \quad \text{for } 1 \leq i, j \leq 20, \; i \neq j.

Prove that BB is nonsingular.

?
Problem 7.9.17

Let

A=(1234). A = \begin{pmatrix} 1 & 2 \\ 3 & 4 \end{pmatrix}.

Show that every real matrix BB such that AB=BAAB = BA has the form sI+tAsI+tA, where s,t∈Rs, t \in \mathbb {R}.

?
Problem 7.9.18

Let AA be a 2×22 \times 2 matrix over C\mathbb {C} which is not a scalar multiple of the identity matrix II. Show that any 2×22 \times 2 matrix XX over C\mathbb {C} commuting with AA has the form X=αI+βAX = \alpha I+\beta A, where α,β∈C\alpha , \beta \in \mathbb {C}.

?
Problem 7.9.19
?
(1)

Determine the commutant of the n×nn \times n Jordan matrix

A=(λ10⋯000λ1⋯0000λ⋯00⋮⋮⋮⋱⋮⋮000⋯λ1000⋯0λ) A = \begin{pmatrix} \lambda & 1 & 0 & \cdots & 0 & 0 \\ 0 & \lambda & 1 & \cdots & 0 & 0 \\ 0 & 0 & \lambda & \cdots & 0 & 0 \\ \vdots & \vdots & \vdots & \ddots & \vdots & \vdots \\ 0 & 0 & 0 & \cdots & \lambda & 1 \\ 0 & 0 & 0 & \cdots & 0 & \lambda \end{pmatrix}

In particular, determine the dimension of the commutant as a complex vector space.

(2)

What is the dimension of the commutant of the 2n×2n2n \times 2n matrix

J⊕J=(J00J)? J \oplus J = \begin{pmatrix} J & 0 \\ 0 & J \end{pmatrix}?
Problem 7.9.20

Let

A=(2−10−12−10−12). A = \begin{pmatrix} 2 & -1 & 0 \\ -1 & 2 & -1 \\ 0 & -1 & 2 \end{pmatrix}.

Show that every real matrix BB such that AB=BAAB = BA has the form

B=aI+bA+cA2 B = aI+bA+cA^{2}

for some real numbers aa, bb, and cc.

?
Problem 7.9.21

A square matrix AA is nilpotent if Ak=0A^{k} = 0 for some positive integer kk.

?
(1)

If AA and BB are nilpotent, is A+BA+B nilpotent?

(2)

Prove: If AA and BB are nilpotent matrices and AB=BAAB = BA, then A+BA+B is nilpotent.

(3)

Prove: If AA is nilpotent then I+AI+A and I−AI-A are invertible.

Problem 7.9.22

Consider the family of square matrices A(θ)A(\theta ) defined by the solution of the matrix differential equation

dA(θ)dθ=BA(θ) \frac{dA(\theta )}{d\theta } = BA(\theta )

with the initial condition A(0)=IA(0) = I, where BB is a constant square matrix.

?
(1)

Find a property of BB which is necessary and sufficient for A(θ)A(\theta ) to be orthogonal for all θ\theta; that is, A(θ)t=A(θ)−1A(\theta )^{t} = A(\theta )^{-1}, where A(θ)tA(\theta )^{t} denotes the transpose of A(θ)A(\theta ).

(2)

Find the matrices A(θ)A(\theta ) corresponding to

B=(01−10) B = \begin{pmatrix} 0 & 1 \\ -1 & 0 \end{pmatrix}

and give a geometric interpretation.

Problem 7.9.23

Let AA be an r×rr \times r matrix of real numbers. Prove that the infinite sum

eA=I+A+A22+⋯+Ann!+⋯ e^{A} = I+A+\frac{A^{2}}{2}+\cdots +\frac{A^{n}}{n!}+\cdots

of matrices converges (i.e., for each i,ji, j, the sum of (i,j)(i,j)th entries converges), and hence that eAe^{A} is a well-defined matrix.

?
Problem 7.9.24

Show that

det⁡(exp⁡(M))=etr⁡(M) \det (\exp (M)) = e^{\operatorname {tr}(M)}

for any complex n×nn \times n matrix MM, where exp⁡(M)\exp (M) is defined as in Problem 7.9.23.

?
Problem 7.9.25

Let TT be an n×nn \times n complex matrix. Show that

lim⁡k→∞Tk=0 \lim _{k \to \infty } T^{k} = 0

if and only if all the eigenvalues of TT have absolute value less than 11.

?
Problem 7.9.26

Let AA and BB be n×nn \times n complex matrices. Prove that

∣tr⁡(AB∗)∣2≤tr⁡(AA∗)tr⁡(BB∗). \left|\operatorname {tr}(AB^{*})\right|^{2} \leq \operatorname {tr}(AA^{*})\operatorname {tr}(BB^{*}).
?
Problem 7.9.27

Let AA and BB be n×nn \times n real matrices, and kk a positive integer. Find

?
(1)

lim⁡t→01t((A+tB)k−Ak)\displaystyle \lim_{t \to 0} \frac{1}{t}\left((A+tB)^{k}-A^{k}\right).

(2)

ddttr⁡(A+tB)k∣t=0\displaystyle \frac{d}{dt}\operatorname {tr}(A+tB)^{k}\bigg|_{t=0}.

Problem 7.9.28
?
(1)

Prove that any real n×nn \times n matrix MM can be written as M=A+S+cIM = A+S+cI, where AA is antisymmetric, SS is symmetric, cc is a scalar, II is the identity matrix, and tr⁡S=0\operatorname {tr} S = 0.

(2)

Prove that with the above notation,

tr⁡(M2)=tr⁡(A2)+tr⁡(S2)+1n(tr⁡M)2. \operatorname {tr}(M^{2}) = \operatorname {tr}(A^{2})+\operatorname {tr}(S^{2})+\frac{1}{n}(\operatorname {tr} M)^{2}.
Problem 7.9.29

Let AA be an n×nn \times n complex matrix such that tr⁡Ak=0\operatorname {tr} A^{k} = 0 for k=1,…,nk = 1, \ldots , n. Prove that AA is nilpotent.

?
Problem 7.9.30

Let NN be a nilpotent complex matrix. Let rr be a positive integer. Show that there is a n×nn \times n complex matrix AA with

Ar=I+N. A^{r} = I+N.
?
Problem 7.9.31

Let A=(aij)i,j=1nA = (a_{ij})_{i,j=1}^{n} be a real n×nn \times n matrix such that aii≥1a_{ii} \geq 1 for all ii, and

∑i≠jaij2<1. \sum _{i \neq j} a_{ij}^{2} < 1.

Prove that AA is invertible.

?
Problem 7.9.32

Show that an n×nn \times n matrix of complex numbers A=(aij)A = (a_{ij}) satisfying

∣aii∣>∑j≠i∣aij∣ \left|a_{ii}\right| > \sum _{j \neq i} \left|a_{ij}\right|

for 1≤i≤n1 \leq i \leq n must be invertible.

?
Problem 7.9.33

Let A=(aij)i,j=1nA = (a_{ij})_{i,j=1}^{n} be an n×nn \times n matrix such that ∑j=1n∣aij∣<1\sum_{j=1}^{n} \left|a_{ij}\right| < 1 for each ii. Prove that I−AI-A is invertible.

?
Problem 7.9.34

Let AA be a real n×nn \times n matrix. Let MM denote the maximum of the absolute values of the eigenvalues of AA.

?
(1)

Prove that if AA is symmetric, then ∥Ax∥≤M∥x∥\left\| Ax\right\| \leq M\left\| x\right\| for all xx in Rn\mathbb {R}^{n}. (Here, ∥⋅∥\left\| \cdot \right\| denotes the Euclidean norm.)

(2)

Prove that the preceding inequality can fail if AA is not symmetric.

Problem 7.9.35

Let AA be an n×nn \times n matrix over C\mathbb {C} whose minimal polynomial has degree kk.

?
(1)

Prove that, if the point λ\lambda of C\mathbb {C} is not an eigenvalue of AA, then there is a polynomial pλp_{\lambda } of degree k−1k-1 such that pλ(A)=(A−λI)−1p_{\lambda }(A) = (A-\lambda I)^{-1}.

(2)

Let λ1,…,λk\lambda_{1}, \ldots , \lambda_{k} be distinct points of C\mathbb {C} that are not eigenvalues of AA. Prove that there are complex numbers c1,…,ckc_{1}, \ldots , c_{k} such that

∑i=1kcipλi=1. \sum _{i=1}^{k} c_{i}p_{\lambda _{i}} = 1.
Problem 7.9.36

Let ∥x∥\left\| x\right\| denote the Euclidean norm of a vector xx. Show that for any real m×nm \times n matrix MM there is a unique non-negative scalar σ\sigma, and (possibly non-unique) unit vectors u∈Rmu \in \mathbb {R}^{m} and v∈Rnv \in \mathbb {R}^{n} such that

?
(1)

∥Mx∥≤σ∥x∥\left\| Mx\right\| \leq \sigma \left\| x\right\| for all x∈Rnx \in \mathbb {R}^{n},

(2)

Mu=σvMu = \sigma v,

(3)

MTv=σuM^{T}v = \sigma u (where MTM^{T} is the transpose of MM).