MATH 33A

SECOND MIDTERM EXAMINATION

May 20th, 2013
Please show your work. You will receive little or no credit for a correct answer to a
problem which is not accompanied by sucient explanations. If you have a question about
any particular problem, please raise your hand and one of the proctors will come and talk to
you. At the completion of the exam, please hand the exam booklet to your TA. If you have
any questions about the grading of the exam, please see the instructor within 15 calendar
days of the examination (i.e. by June 4th).
I certify that the work appearing on this exam is completely my own:
Signature:
Name: Section:
#1 #2 #3 #4 #5 Total
R
1
2 33A MIDTERM 2 RADKO
Problem 1. Let V be the subspace of R
4
consisting of vectors
_

_
x
1
x
2
x
3
x
4
_

_
such that
x
1
+ x
2
+ x
3
+ x
4
= 0
x
1
x
2
+ x
3
+ x
4
= 0
(1) Find a basis for V .
Solution: Notice that
V = ker(T),
where T : R
4
R
2
is a linear transformation whose matrix in the standard basis is
A = [T] =
_
1 1 1 1
1 1 1 1
_
.
The rank of this matrix is equal to 2. (It can not be more than 2 since we only
have two rows. It can not be less than 2 since the rst two columns are linealy
independent).
By rank-nullity theorem, the nullity of this matrix is 4 2 = 2. Thus, the vector
space V is two dimensional.
Since the rst and the third columns are identical, the vector
_
_
_
_
1
0
1
0
_
_
_
_
is in V . Similarly, since the fourth and the third columns are identical, the vector
_
_
_
_
0
0
1
1
_
_
_
_
is in V . These two non zero vectors are linearly independent. Since V is two dimen-
sional, they form a basis of V .
(2) Find the dimension of V .
Solution: see above.
(3) Find an orthonormal basis for P.
33A MIDTERM 2 Radko 3
Solution: apply the Gramm-Schmidt to the vectors v
1
and v
2
above:
u
1
=
1

2
_
_
_
_
1
0
1
0
_
_
_
_
v

2
= v
2
(v
2
u
1
)u
1
=
_
_
_
_
1
2
0
1
2
1
_
_
_
_
u
2
=
_
_
_
_
_
1

6
0
1

_
2
3
_
_
_
_
_
=
1

6
_
_
_
_
1
0
1
2
_
_
_
_
.
The vectors u
1
and u
2
form an orthonormal basis.
4 33A MIDTERM 2 RADKO
Problem 2. (10 pts) Let B be the basis for R
2
given by the vectors v
1
=
_
1
2
_
, v
2
=
_
3
4
_
.
(1) Let w
1
=
_
2
2
_
and w
2
=
_
2012
2013
_
be two vectors in R
2
. Find the coordinates
B
[w
1
]
and
B
[w
2
] of w
1
and w
2
in the basis B.
Solution: By inspection, notice that
v
1
+ w
1
= v
2
.
Thus,
B
[w
1
] =
_
1
1
_
.
To nd coordinates of w
2
in this basis, form the matrix of the change of basis:
S =
_
1 3
2 4
_
.
Then
_
2012
2013
_
=
_
1 3
2 4
_

_
x
y
_
,
where x, y are the coordinates we are looking for. Finding rref, we get
rref
_
1 3 | 2012
2 4 | 2013
_
=
_
1 0 |
2009
2
0 1 |
2011
2
_
.
Thus,
B
[w
2
] =
_

2009
2
2011
2
_
.
(2) Let T be the transformation of R
2
given by the matrix A = [T] =
_
7 3
8 3
_
in the
standard basis. Find the matrix
B
[T] of T in the basis B.
Solution:
B
[T] =
_
_
| |
B
[Tv
1
]
B
[Tv
2
]
| |
_
_
Since Tv
1
=
_
1
2
_
, Tv
2
=
_
9
12
_
in the standard basis, we get
B
[Tv
1
] =
_
1
0
_
,
B
[Tv
2
] =
_
0
3
_
.
33A MIDTERM 2 Radko 5
Thus,
B
[T] =
_
1 0
0 3
_
.
6 33A MIDTERM 2 RADKO
Problem 3. (10 points)
Let I be the 5 5 identity matrix. Is there a 5 5 matrix A for which A
T
A = I
(here A
T
denotes the transpose of A)? Either construct an example of such a matrix A,
or explain why such a matrix can not exist. (The entreis of A can be arbitrary real numbers).
Solution:
Method 1: Since det(A B) = det(A) det(B) and det(A
T
) = det(A), it follows that
A
T
A = I implies (det(A))
2
= 1. Since det(A) is a real number, this is not possbile.
Thus, there is no such matrix.
Method 2: Let b
11
be the left upper entry in the matrix A
T
A. The left upper entry
in I is 1. Compute b
11
in terms of entries of A:
b
11
= a
2
11
+ a
2
21
+ + a
2
n1
.
Since the sum of the squares is non-negative, it can not equal to 1. Thus, no such matrix
could possibly exist.
33A MIDTERM 2 Radko 7
Problem 4. (10 points) Find the least squares solution x

of the system Ax =

b, where
A =
_
_
1 1
1 0
0 1
_
_
, b =
_
_
3
3
3
_
_
.
Solution: To start solving the problem, compute A
T
A =
_
2 1
1 2
_
and A
T
b =
_
6
6
_
.
Then
x = (A
T
A)
1
A
T
b =
1
3
_
2 1
1 2
_

_
6
6
_
=
_
2
2
_
.
8 33A MIDTERM 2 RADKO
Problem 5. (True/False, 1 pt each) Mark your answers by lling in the appropriate box
next to each question.
(a
T F
) Let A be an orthogonal 4 4 matrix such that Ae
1
= e
2
, Ae
2
= e
3
and Ae
3
= e
1
.
Then Ae
4
= e
4
.
False: Ae
4
= e
4
also works.
(b
T F
) Let {x, y, z} be a set of linear independent vectors in a linear subspace V of R
n
.
Then no two vectors u, w V can span V .
True: if there are 3 linearly independent vectors, the dimension of V is 3.
Thus, no two vectors can span it.
(c
T F
) If A is an orthogonal matrix, all the entries of A are less than or equal to 1.
True: columns of an orthogonal matrix are images of the standard basis vec-
tors under an orthogonal transformations. Since an orthogonal transformation
preserves lengths, the columns represent vectors of length 1. Therefore, their
components can not be bigger than 1.
(d
T F
) The dimension of the vector space of all symmetric 4 4 matrices is 10.
True: This vector space is spanned by 10 symmetric matrices of the form
(E
ij
+E
ji
), where 1 i j 4 and E
ij
denotes the matrix which has entry a
ij
= 1
and the rest of the entries are 0s.
(e
T F
) Let T : R
5
R
3
be a linear transformation such that ker(T) has dimension 2. Then
dimension of the image(T) is equal to 1.
False: By rank-nullity theorem, nullity(T) + dim(im(T)) = 5. Thus, if nullity
is 2, the dimension of the image is 3.
(f
T F
) If determinant of a matrix is equal to 1, the matrix is similar to the identity matrix.
False: Consider the following counterexample:
_
2 0
0
1
2
_
. While the determi-
nant of this matrix is 1, it can not be similar to the identity matrix since its trace is
2
1
2
= 2 (the trace of the identity 2 2 matrix).
(g
T F
) Orthogonal projection is an example of an orthogonal transformation.
False: Orthogonal transformation preserve length (by denition). An orthogo-
nal projection does not preserve length.
33A MIDTERM 2 Radko 9
(h
T F
) A linear transformation which takes an orthonormal basis into another orthonormal
basis is orthogonal.
True: Since the transformation preserves the lengths of the basis vectors, it
will preserve the lengths of all vectors.
(i
T F
) For any invertible matrix we have (A
T
)
1
= (A
1
)
T
.
True: A A
1
= I implies (A A
1
)
T
= (A
1
)
T
A
T
= I. Thus, (A
1
)
T
= (A
T
)
1
.
(j
T F
) For any matrix A one can nd Q and R such that A = QR, where Q is an orthogonal
matrix and R is an upper triangular matrix with non-negative entries on the diagonal.
False: If A is not a square matrix, it is not true.