Scribd
Upload
32 views4 pages

NRNXN

The document is an exam for a mathematics course on matrix theory and its applications. It contains 6 questions testing various concepts in matrix theory, including properties of similar matrices, eigenvalues and eigenvectors, idempotent matrices, pseudo-inverses, matrix norms, and normal matrices. The exam instructs students to show their work and provide clear explanations for their conclusions. It also provides hints for some questions and specifies the point values for each question.

Uploaded by

Wai man Law
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
32 views4 pages

NRNXN

The document is an exam for a mathematics course on matrix theory and its applications. It contains 6 questions testing various concepts in matrix theory, including properties of similar matrices, eigenvalues and eigenvectors, idempotent matrices, pseudo-inverses, matrix norms, and normal matrices. The exam instructs students to show their work and provide clear explanations for their conclusions. It also provides hints for some questions and specifies the point values for each question.

Uploaded by

Wai man Law
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

THE UNIVERSITY OF HONG KONG

DECEMBER 2008 EXAMINATION

MATHEMATICS: PAPER MATH2303 MATRIX THEORY AND ITS


APPLICATIONS

(To be taken by BSc, BEcon, BEcon&Fin, BEng( CompSc), BEng(EComE),


BSc( ActuarSc) & MPhil students)

12 December, 2008 9:30a.m. - 12:00noon

Candidates may use any self-contained, silent, battery-operated and pocket-sized cal-
culator. The calculator should have numerical-display facilities only and should be
used only for the purposes of calculation. It is the candidate's responsibility to ensure
that his/her calculator operates satisfactorily.

Candidates must record the name and type of their calculators on the front page of
their examination scripts.

Answer ALL SIX Questions

Note: You should always give precise and adequate explanations to support your
conclusions. Clarity of presentation of your argument counts. So think carefully
before you write.

1. (15 points)

Let A, B E nrnxn.

(a) Show that AB is similar to BA if either A orB is nonsingular.

(b) Show that (A: ~:) is similar to ( ~ ~~) . Hence, or otherwise, show
that AB and BA have the same characteristic polynomial.

.
[H1nt: Cons1.der (AB On) (In A) and (In A) (On OnA) . Here On
B On On In On In B B
and In denote then x n zero and identity matrices respectively.]

(c) Give an example of A, BE R 2 x 2 in which AB is not similar to BA. Explain


your answer.

1
2. (15 points)

Let A E JFnxn. Suppose that A is an eigenvalue of A with algebraic multiplicity


ma and geometric multiplicity m 9 • Prove or disprove the following:

(a) A is also an eigenvalue of AT with algebraic multiplicity ma and geometric


multiplicity m 9 .

(b) A and AT have the saine eigenvectors.

(c) For all positive integers k, the algebraic multiplicity of )..k as an eigenvalue
of Ak is also ma.

3. (15 points)

Suppose A is a square matrix of order n which is idempotent, that is, A satisfies


A 2 =A.

(a) Show that the eigenvalues of A are either 0 or 1.

(b) Show that A is similar to In-k EB Ok for 0 :5 k :5 n, where In-k and Ok


denote respectively the (n- k) x (n- k) identity matrix and the k x k zero
matrix. [Hint: Consider the possible Jordan form of A.]

4. (20 points)

(a) Let A E JFmxn and bE JFm. Prove that the vector x :=A-b, where A- is
the pseudo-inverse of A, is the unique vector in JFn that satisfies

(i) I!Ax - bilE :5 IlAx- bilE for all x E JFn, and

(ii) llxiiE < llxiiE for all x E JFn which satisfy IIAx- bilE= IIAx- bilE and
x # x. (Here 11· liE denotes the Euclidean norm).

(b) Let A E JFmxr and B E JFrxn. Show that if rankA = rankB = r, then
(AB)-=B-A-.

2
5. (15 points)

Let A= (aii) E cmxn have rank r and nonzero singular values 0"1 ~ ••• ~ O"r.

Denote by 11 · IIF and 11 · ll2 respectively the Frobenius norm and the spectral
norm, defined by IIAIIF = Jtr (A* A) and IIAII2 = o-1.

(a) Show that


m n r
(i) IIAII} = L L laii1 = L O";, 2

i=1 j=1 i=1

(b) Show that IIAIIF = IIAII2 if and only if rank A < 2.

6. (20 points)

(a) Let A E lFmxn have rank r ~ 1 and nonzero singular values 0"1 ~ ··• ~ O"r.

Define a (m+ n) x (m+ n) matrix

H(A) = (~: ~),


where Om and On are m x m and n x n zero matrices respectively.

(i) Show that H(A) is Hermitian.


(ii) Show that the m+ n eigenvalues (counting algebraic multiplicities) of
H(A) are

[Hint: Consider the singular value decomposition (version 2) of A and


try to write H (A) in the form of W DW* where W has orthonormal
columns and D is a diagonal matrix.]

(b) Let A E cnxn. Define AH = ~(A+ A*), the Hermitian part, and AK =
~(A- A*), the skew-Hermitian part of A. Then A= AH + AK. Show that

the following statements are equivalent:

(i) A is normal.
(ii) AH and AK commute with each other.

3
(iii) AH and AK. are simultaneously diagonalizable by unitary similarity.
[X, Y E cnxn are simultaneously diagonalizable by unitary similarity
if there exists a unitary matrix U E cnxn such that both U* XU and
U*YU are diagonal.]

* * * * * * * End of Paper * * * * * * *

You might also like