and
1
N
or
1
MN
, which
are not equal. Demonstrate by calculation that
1 1
N M MN
= using the following:
2 1
0 1
M
(
=
(
and
1 0
3 2
N
(
=
(
.
Warning: In this class, it is a tradition that you will get 0 point for an exam, quiz, or
homework problem if you use the invalid notation such as
M
N
regardless of what work
you have done for the problem.
Problem 7. Let T and A be n n matrices and suppose T is non-singular. Show that
1 1
( )
k k
T AT T A T
= .
(Hint: For an n n matrix M,
k
M MM M = , i.e., the product of k Ms.)
Instructor: Hamano
EE511 HW#1 Fall 2013, CSULB 2013 Fumio Hamano Page 3 of 6
Problem 8. A square matrix M is said to be orthogonal if ' M M I = , i.e,
1
M M
' = .
Here, ' M is the transpose of M and I is an identity matrix of an appropriate size.
Answer the following questions.
(1) Show that det I =1.
(2) Show that, if matrix M is orthogonal, then det 1 M = . (Hint: Take the
determinant of the both sides of the equation ' M M I = .)
Problem 9. Let M, N, A, and W be n n nonsingular matrices. Also, let
n
I be an n n
identity matrix. Suppose | | | |
n n
MN A I I W = . Answer the following questions.
(1) Write 2 separate matrix equations from the above equation.
(2) Show that
1
W A
=
(3) Show that ( ) det 1 det det A M N = .
Problme10. Let
1
C ,
2
C ,
11
A ,
12
A ,
22
A , and
1
B be respectively
1
p n ,
2
p n ,
1 1
n n ,
1 2
n n ,
2 2
n n , and
1
n m matrices and let s be a scalar. Also, let I and 0 be identity and
null (or zero) matrices of appropriate dimensions, respectively. Show
| | ( )
1
1
11 12 1
1 2 1 11 1
22
0 0
sI A A B
C C C sI A B
sI A
( (
=
( (
.
Hint: Use
1
1 1 1
11 12 11 11 12 22
1
22 22
0 0
M M M M M M
M M
( (
=
( (
.
Problem 11. Let A, B, and C be respectively, n n , n m , and p n matrices. Calculate
the product
1
1
n
n
C
CA
B AB A B
CA
(
(
(
(
(
(
.
Note: The resulting matrix is called Hankel matrix.
Problem 12. Show
1 1 2 3 2
( ) sI A s I s A s A
= + + + .
Hint: Note the above equality holds iff
1 2 3 2
( )( ) I sI A s I s A s A
= + + + . Show that
this equation holds.
Problem 13. Let A, B, and F be n n , n m , and m n matrices, respectively, and let I
be an identity matrix of appropriate dimension. Also, let s be a scalar. Show that,
assuming all inverses are well-defined,
{ }
1
1 1
( ) ( ) I F sI A BF B I F sI A B
+ = .
Instructor: Hamano
EE511 HW#1 Fall 2013, CSULB 2013 Fumio Hamano Page 4 of 6
Hint: Show the above equality holds iff
{ }{ }
1 1
( ) ( ) I F sI A BF B I F sI A B I
+ = .
After you expand the left hand side of this equality, replace the term
1
( ) F sI A BF B
with its equivalent
1 1
( ) ( )( ) F sI A BF sI A sI A B
.
Derivative of Matrices:
Problem 14. Let , , and A B C be respectively , , and m m n n p matrices whos
elements are functions of t. Show that the following equalities hold:
(1) ( )
d
AB AB AB
dt
= +
In particular, if A is a constant matrix, then ( )
d
AB AB
dt
= , and if B is a constant matrix,
then ( )
d
AB AB
dt
= .
(2) ( )
d
ABC ABC ABC ABC
dt
= + + .
Gradient and Jacobian:
Problem 15. Let
1 1 1 2 1 1 2
( ) ( , ) f x f x x x x x = = +
2 2 1 2 2 1 2
( ) ( , ) f x f x x x x x = =
where | |
1 2
: x x x
'
= .
Answer the following questions.
(1) Find the gradient
1
( ) f x
x
c
c
of
1
( ) f x . (Represent it as a row vector.)
(2) Find the Jacobian
1
2
( )
( )
f x
f x x
( c
(
c
.
Eigen-Values and Eigen-Vectors:
Review:
Let A be an n n (real or complex) matrix and suppose
Av v =
for some C e and a non-zero
n
v C e . Then, is called an eigen-value of A and v is
called an eigen-vector of A corresponding to the eigen-value . From the equation above,
we must have
( ) det 0 I A = .
This equation can be used to determine the eigen-values of A.
Instructor: Hamano
EE511 HW#1 Fall 2013, CSULB 2013 Fumio Hamano Page 5 of 6
Fact. If A has n distinct eigen-values
1
, ,
n
, then the set of n eigen-vectors
corresponding to the above eigen-values is linearly independent.
Problem 16. Let A be a matrix given by
0 1 0
0 0 1
6 11 6
A
(
(
=
(
(
.
(Note: A matrix having 1s on the super-diagonal and possibly non-zero elements on the
bottom row, and otherwise having zeros all over is said to be in a companion form. A
form of matrix, which is the transpose of the above, is also called a companion form.)
Answer the following questions.
(1) Find the eigen-values
i
, i=1, 2, 3 and corresponding eigen-vectors
i
v , i=1, 2, 3
of A.
(2) If we define | |
1 2 3
: T v v v = , what is
1
T AT
equal to?
Note: An operation of the form
1
(.) T T
= .
Problem 17. Let A be a 2 2 real matrix, which has eigen-values j o e + and j o e
with 0 e = and
2
1 j = and the corresponding eigen-vectors
1
e and
2
e , respectively. Let
R
e and
Im
je be the real and imaginary parts of
1
e . Define
: M
o e
e o
(
=
(
and | |
Im
:
R
T e e =
and show
1
T AT M
= , or equivalently, AT TM =
Note: In linear algebraic terms, the above means that, with respect to the basis
{ }
Im
,
R
e e , the map A has its (matrix) representation
: M
o e
e o
(
=
(
.
Problem 18. Let A be an n n nonsingular matrix. Show that, if is an eigen-value of A,
then1 is an eigen-value of
1
A
.
Problem 19. Let A be an n n real matrix. Show that, if a complex number is an
eigen-value of A, then its complex conjugate * is also an eigen-value of A.
Furthermore, if v is an eigen-vector corresponding to the eigen-value , then its complex
conjugate * v is an eigen-vector corresponding to * .
Instructor: Hamano
EE511 HW#1 Fall 2013, CSULB 2013 Fumio Hamano Page 6 of 6
Problem 20. Let A be an n n matrix. Show that the set of eigen-values does not change
with similarity transformation of A, i.e., the eigen-values of A are identical to those of
1
: A T AT