Math 2040: Matrix Theory and Linear Algebra II

Solutions to Assignment 2

Section 4.5 Dimension

4.5.6. Problem
 Restatement:
 Find (a) a basis, and (b) state the dimension of the subspace
3a + 6b − c
 6a − 2b − 2c 
{ 
 −9a + 5b + 3c  ; a, b, c ∈ R}.
−3a + b + c
   
3 6
 6   −2 
Final Answer: (a) {    3
 −9  ,  5 } is a basis of the R subspace.
−3 1
(b) The dimension of the subspace is 2 since a basis for it has two vectors in it.
       
3a + 6b − c 3 6 −1
 6a − 2b − 2c       
Work: (a) {  ; a, b, c ∈ R} = Span{ 6  ,  −2  ,  −2 }. The first
 −9a + 5b + 3c   −9   5   3 
−3a + b + c −3 1 1
and third vectors are multiples of each other, and the first and second vectors are linearly
independent of each other so we can remove the third to get a basis (that is, a linearly
independent set of spanning vectors) of the subspace.
(b) No work required.
   Determine
4.5.12. Problem Restatement:   the
 dimension
 of
1 −3 −8 −3
W = Span{ −2  ,  4  ,  6  ,  0 }.
0 1 5 7
Final Answer: dim(W ) = 3.
   
1 −3 −8 −3 1 −3 −8 −3
Work: dim(W ) = dim(Col  −2 4 6 0 ) = dim(Col  0 1 5 3 ).
0 1 5 7 0 0 0 1
Since there are 3 pivot columns, dim(W ) = 3.
4.5.14. Problem Restatement: Determine the dimensions of N ul(A) and Col(A) for

1
  
1 3 −4 2 −1 6
 0 0 1 −3 7 0 
A=
 0
.
0 0 1 4 −3 
0 0 0 0 0 0
Final Answer: dim(Col(A)) = Rank(A) = 3 and dim(N ul(A)) = 6 − Rank(A) = 6 − 3 = 3.
Work: A is in echelon form and we count 3 pivot columns in A.
4.5.24. Problem Restatement: Let B be the basis of P2 consisting of the first three Laguerre
polynomials, {1, 1 − t, 2 − 4t + t2 }, and let p(t) = 7 − 8t + 3t2 . Find the coordinate vector
of p relative to B.
 
5
Final Answer: [p(t)]B =  −4 .
3
Work: Several equivalent methods to solve this problem are possible. One of thesefollows. 
7
Let S = {1, t, t2 } be the standard basis of P2 . [p(t)]B = PB−1 (p(t)) = PB−1 (PS  −8 ) =
3
     
7 7 7
(PB−1 PS )  −8  = (PS−1 PB )−1  −8 . Therefore, if b =  −8 , then [p(t)]B is the solu-
3 3 3
 
1 1 2
tion of (PS−1 PB )x = b. Now PS−1 PB is easy to construct; it is the matrix  0 −1 −4 .
0 0 1
Passing to the  augmented matrix and
  row reducing toreduced echelon form gives us
1 1 2 | 7 1 0 0 | 5 5
[PS−1 PB |b] =  0 −1 −4 | −8  ∼  0 1 0 | −4 . Therefore, [p(t)]B =  −4 .
0 0 1 | 3 0 0 1 | 3 3
4.5.26. Problem Restatement: Let H be an n-dimensional subspace of an n-dimensional vector
space V . Show that H = V .
Final Answer: Choose a basis B = {b1 , ..., bn } of H, so that H = Span(B). By the Basis
Theorem, page 257, B is also a basis of V , so V = Span(B). Therefore, H = V , since they
are each equal Span(B).
Work: None required.

2
 Section 4.6 Rank

4.6.2. Problem
 Restatement: Assumethat A is row  equivalent to B where
1 −3 4 −1 9 1 −3 0 5 −7
 −2 6 −6 −1 −10   8 
A=  and B =  0 0 2 −3 .
 −3 9 −6 −6 −3   0 0 0 0 5 
3 −9 4 9 0 0 0 0 0 0
Without calculation, list Rank(A) and dim(N ul(A)). Then find a basis for Col(A), Row(A)
and N ul(A).
Final Answer:
Rank(A) = 3 and dim(N ul(A)) = 2.
A basis of Col(A) is given by the first, third, and fifth columns of A.
A basis of Row(A) is given by the first three rows of A (or of B).
   
3 −5
 1   0 
   
A basis of N ul(A) is { 0  ,  3/2 } (see work below).
   
 0   1 
0 0
 
1 −3 0 5 0
 0 0 1 −3/2 0 
Work: A ∼ B ∼   0
, so if x is a solution of the homogeneous system
0 0 0 1 
0 0 0 0 0
Ax = 0, we have x2 and x4 free, giving us
x1 = 3x2 − 5x4 ,
x2 = x1 + 0x4 ,
x3 = 0x2 + (3/2)x4 ,
x4 = 0x2 + x4 ,
x5 = 0x2 + 0x4 .
4.6.6. Problem Restatement: If a 6 × 3 matrix A has rank 3, find dim(N ul(A)), dim(Row(A)),
and Rank(AT ).
Final Answer: A : R3 → R6 , A having 6 rows and 3 columns. dim(N ul(A)) = 3−Rank(A) =
3 − 3 = 0, by the Rank Theorem. dim(Row(A)) = Rank(A), so dim(Row(A)) = 3, and
Rank(AT ) = Rank(A) so Rank(AT ) = 3.
Work: None required.

3
4.6.10. Problem Restatement: If the null space of a 7 × 6 matrix A is 5-dimensional, what is the
dimension of the column space of A?
Final Answer: The dimension of the column space of A is Rank(A) by definition. Then, by
the Rank Theorem, we have dim(Col(A)) = 6 − dim(N ul(A)) = 6 − 5 = 1.
Work: None required.
4.6.16. Problem Restatement: If A is a 6 × 4 matrix, what is the smallest possible dimension of
N ul(A)?
Final Answer: Zero is the smallest possible dimension of the null space of a 6 × 4 matrix A.
Work: A : R4 → R6 , A having 6 rows and 4 columns. 0 ≤ dim(N ul(A)), but if
{e1 , e2 , e3 , e4 , e5 , e6 } is the standard basis of R6 and A = [e1 e2 e3 e4 ], say, then A is one to
one and, therefore, N ul(A) = {0}, so dim(N ul(A)) = 0. Therefore, if A is a 6 × 4 matrix,
the smallest possible dimension of N ul(A) is 0.
4.6.30. Problem Restatement: Suppose A is m×n and b is in Rm . What has to be true about the
two numbers Rank([A b]) and Rank(A) in order for the equation Ax = b to be consistent?
Final Answer: It must be that Rank([A b]) = Rank(A).
Work: In order for Ax = b to be consistent, the last column of the augmented matrix [A b]
cannot be a pivot column. Therefore, in order for Ax = b to be consistent, Rank([A b])
must equal the number of pivot columns of A.
Section 4.7 Change of Basis

4.7.2. Problem Restatement: Let B = {b1 , b2 } and C = {c1 , c2 } be bases for a vector space V ,
and suppose b1 = −c1 + 4c2 and b2 = 5c1 − 3c2 .
a. Find the change-of-coordinate matrix from B to C.
b. Find [x]C for x = 5b1 + 3b2 .
· ¸ · ¸
−1 5 10
Final Answer: a. PC←B = . b. [x]C = .
4 −3 11
Work: a. PC←B = PC−1 PB = [PC−1 P·B e1 PC−1 PB¸e2 ] = [PC−1 b1 PC−1 b2 ] = [[b1 ]C [b2 ]C ]
−1 5
= [[−c1 + 4c2 ]C [5c1 − 3c2 ]c ] = .
4 −3
· ¸ · ¸ · ¸ · ¸
5 −1 5 10
b. [x]C = [5b1 + 3b2 ]C = 5[b1 ]C + 3[b2 ]C = PC←B =5 +3 = .
3 4 −3 11

4
4.7.6. Problem Restatement: Let D = {d1 , d2 , d3 } and F = {f1 , f2 , f3 } be bases for a vector
space V , and suppose f1 = 2d1 − d2 + d3 , f2 = 3d2 + d3 , and f3 = −3d1 + 2d3 .
a. Find the change-of-coordinate matrix from F to D.
b. Find [x]D for x = f1 − 2f2 + 2f3 .
   
2 0 −3 −4
Final Answer: a. PD←F =  −1 3 0 . b. [x]D =  −7 .
1 1 2 3
Work: See 4.7.2. above for methodology, replacing B by F and C by D throughout and
adjusting for dimension.
· ¸ · ¸ · ¸ · ¸
−1 1 1 1
4.7.8. Problem Restatement: Let B = { , } and C = { , } be bases for
8 −5 4 1
R2 . Find the change-of-coordinated matrix from B to C and the change-of-coordinated
matrix from C to B. · ¸ · ¸
3 −2 3 2
Final Answer: PC←B = and PB←C = .
−4 3 4 3
· ¸−1 · ¸ · ¸· ¸
−1 1 1 −1 1 1 1 −1 −1 1
Work: PC←B = PC PB = = −3
4 1 8 −5 −4 1 8 −5
· ¸ · ¸
3 −2 −1 3 2
= . PB←C = PC←B = . Other calculations will lead to the same
−4 3 4 3
answers.
4.7.14. Problem Restatement: In P2 , find the change-of-coordinates matrix from the basis B =
{1 − 3t2 , 2 + t − 5t2 , 1 + 2t} to the standard basis. Then write t2 as a linear combination of
the polynomials in B. 
1 2 1
Final Answer: PS←B =  0 1 2 . t2 = 3(1 − 3t2 ) − 2(2 + t − 5t2 ) + (1 + 2t).
−3 −5 0
Work: The standard basis is S = {1, t, t2 }.
−1 −1 −1 −1 −1 2 −1 2 −1
PS←B
 = PS PB = [PS PB e1 PS PB e2 PS PB e3 ] = [PS (1−3t ) PS (2+t−5t ) PS (1+2t)]
1 2 1
= 0 1 2 .
−3 −5 0
 
0
[t2 ]B = PB−1 (t2 ) = PB−1 PS PS−1 (t2 ) = PB−1 PS  0 
1
 
0
= (PS−1 PB )−1  0  = PS←B −1
e3 .
1

5
 Therefore, [t2 ]B is the solution of the equation PS←B x = e3 . Passing to the augmented
matrix and row reducing, gives us   
1 2 1 | 0 1 0 0 | 3
[PS←B | e3 ] =  0 
1 2 | 0 ∼ ... ∼ 0  1 0 | −2 .
−3 −5 0 | 1 0 0 1 | 1
4.7.x. Problem Restatement: Calculate the change of coordinate matrix PC←B where
B = {1 − 2t + t2 , 3 − 5t + 4t2 , 2t + 3t2 } and C = {1 − 3t2 , 2 + t − 5t2 , 1 + 2t}.
You may assume B and C are each basis of P2 .
 
23 67 −1
Final Answer: PC←B =  −14 −41 0 
6 18 1
Work: PC←B = PC−1 PB = PC−1 PS PS−1 PB = (PS−1 PC )−1 (PS−1 PB ) = PS←C
−1
PS←B . Therefore,
PC←B is the solution of the equation PS←C X = PS←B . The matrices PS←C and PS←B are

1 3 0
easy to formulate (see 4.7.14. above). They are PS←B =  −2 −5 2  and
1 4 3
 
1 2 1
PS←C =  0 1 2 . Passing to the augmented matrix and row reducing allows us to
−3 −5 0
solve for X = PC←B .  
1 2 1 | 1 3 0
[PS←C | PS←B ] =  0 1 2 | −2 −5 2 
−3 −5 0 | 1 4 3
 
1 0 0 | 23 67 −1
∼ ... ∼  0 1 0 | −14 −41 0  = [I3 | PC←B ].
0 0 1 | 6 18 1

6
 Section 5.1 Eigenvectors and Eigenvalues

 
1 2 2
5.1.8. Problem Restatement: Is λ = 3 an eigenvalue of A =  3 −2 1 ? If so, find one
0 1 1
corresponding eigenvector.
 
3
Final Answer: λ = 3 is an eigenvalue of A. A corresponding eigenvector is  2 .
1
   
−2 2 2 1 0 −3
Work: A − λI =  3 −5 1  ∼  0 1 −2 .
0 1 −2 0 0 0
· ¸
10 −9
5.1.10. Problem Restatement: Let A = . Find a basis of the eigenspace of
4 −2
A corresponding to the eigenvalue λ = 4.
· ¸ · ¸
3/2 3
Final Answer: { } or, equivalently, { }.
1 2
· ¸ · ¸
6 −9 1 −3/2
Work: A − λI = ∼ .
4 −6 0 0
 
3 0 2 0
 1 3 1 0 
5.1.16. Problem Restatement: Let A =  
 0 1 1 0 . Find a basis of the eigenspace of
0 0 0 4
A corresponding to the eigenvalue λ = 4.
   
2 0
 3   0 
Final Answer: {   
 1  ,  0 }.
0 4
   
−1 0 2 0 1 0 −2 0
 1 −1 1 0   
Work: A − λI =   ∼  0 1 −3 0 .
 0 1 −3 0   0 0 0 0 
0 0 0 0 0 0 0 0
5.1.24. Problem Restatement: Construct an example of a 2 × 2 matrix with only one distinct
eigenvalue.

7
 Final Answer: By Theorem 1 of section 5.1, we can take any 2 × 2 triangular matrix with
two equal diagonal elements. One example, therefore, is the 2 × 2 zero matrix. Another
example is the 2 × 2 identity matrix I2 . Many other examples are possible.
Work: None required.
5.1.26. Problem Restatement: Show that if A2 is the zero matrix, then the only eigenvalue of A
is 0.
Final Answer: Suppose A2 = 0. We must show 0 is an eigenvalue of A, and we must show
that if λ is an eigenvalue of A then λ = 0.

First, we show 0 is an eigenvalue of A. To do this, in Rn , choose any vector v 6= 0 (tacitly,

we are assuming A is an n × n matrix with n > 0). Let Av = b. There are two cases to
consider; b = 0 and b 6= 0. In the first case, if b = 0 then b = 0v, and this gives us Av = 0v
with v 6= 0, allowing us to conclude 0 is an eigenvalue of A. In the other case, we assume
b 6= 0. Then because A2 = 0, we get 0 = Ab upon multiplying on the left both sides of
Av = b by A. But 0b = 0, so we get 0b = Ab with b 6= 0, allowing us to conclude again that
0 is an eigenvalue of A. Therefore, in both cases (b = 0 and b 6= 0), 0 is shown to be an
eigenvalue of A. This concludes the first part of the proof.

Next, we show that if λ is an eigenvalue of A then λ = 0. To do this, assume λ is an

eigenvalue of A. Therefore, there is a vector v 6= 0 such that Av = λv. Then A2 v =
A(λv) = λAv = λλv = λ2 v. But then 0 = λ2 v because A2 v = 0v = 0. Thus, 0 = λ2
because v 6= 0. Finally, because λ is a scalar, 0 = λ2 implies 0 = λ as required, and this
completes the second part of the proof.
Work: None required.