2.

[16; 8 points each part] Let A be the matrix

 
1 5 0 0 0
0 1 0 0 0
A= 0

First Test Answers 0 0 1 7
Math 130 Linear Algebra 0 0 0 2 14
D Joyce, November 2012
that describes a transformation R5 → R4 .
Scale. 85–100 A, 70–84 B, 50–69 C, 40–49 D. Me- a. Determine the rank and the nullity of A.
dian 72. The fourth row is twice the third, but the first
three are independent since their leading 1’s occur
1. [16] Compute the inverse A−1 of the 3 × 3 ma- in different columns. Therefore the rank is 3. Since
trix   the rank plus the nullity equals 5, therefore the nul-
2 3 1 lity is 2.
A = 1 1 0 . b. Recall that the kernel, also called the null
0 1 2 space, of a matrix A is the set of all solutions of the
The easiest way is to use the algorithm where homogeneous system Ax = 0. Find a basis for the
you write down the 3 × 6 matrix [A|I] and use row kernel A.
operations to convert it to reduced echelon form. As the fourth row is redundant, we can ignore it.
The particular operations you use are up to you. If we denote the 5 variables as x1 through x5 , then
Here’s one sequence of operations. I’ll begin by the three equations say x1 + 5x2 = 0, x2 = 0, and
exchanging the first and second rows, then pivot x4 + 7x5 = 0. Therefore, the general solution is
on the upper left entry, then the middle entry, then
the lower right entry. (x1 , x2 , x3 , x4 , x5 ) = (0, 0, x3 , −7x5 , x5 )
 
2 3 1 1 0 0 where x3 and x5 are freely chosen. Choosing one of
 1 1 0 0 1 0  x3 and x5 to be 1 and the other to be 0, we get two
0 1 2 0 0 1 basis vectors for the kernel, namely
 
1 1 0 0 1 0 (0, 0, 1, 0, 0), and (0, 0, 0, −7, 1).
 2 3 1 1 0 0 
0 1 2 0 0 1 There are, of course, many other pairs of vectors
  that could be chosen as a basis for the kernel. You
1 1 0 0 1 0 could also display them as column vectors instead
 0 1 1 1 −2 0 
of 5-tuples if you prefer.
0 1 2 0 0 1
 
1 0 −1 −1 3 0 3. [28; 4 points each] True/false.
 0 1 1 1 −2 0 
a. If T : V → W is a linear transformation, then
0 0 1 −1 2 1
  T carries linearly independent subsets of V onto
1 0 0 −2 5 1 linearly independent subsets of W . False. T can
 0 1 0 2 −4 −1  send the independent vectors to the same vector,
0 0 1 −1 2 1 even to 0.
Since the left half of the matrix is now a 3 × 3 iden- b. The only square matrices A whose squares
−1
tity matrix, therefore the right half displays A . are the identity, A2 = I, are the two matrices A =

1
I and A = −I. False. There are many others. 5. [24; 4 points each] Identity or not identity.
Every reflection across a line through the origin has
a. A+(BC)D = B(CD)+A. Identity. Addition
a matrix with this property.
is commutative, and multiplication is associative.
c. The set of all linear transformations V → W
b. (A + B)(A − B) = A2 − B 2 . Not identity. The
between vector spaces over a field F forms another
left side is A2 − AB + BA − B 2 , but AB does not
vector space over F . True. If V ∼ = F n and W ∼ =
usually equal BA.
F m , then we can identify this vector space with the
vector space of m × n matrices over F . c. (At + B)t = A + B t . Identity. (At + B)t =
t t t t
d. The vector space M (R) of 3 × 4 matrices (A ) + B = A + B .
3,4
with entries in R is isomorphic to P11 (R), the vec- d. (AB)−1 = A−1 B −1 . Not identity. (AB)−1
tor space of polynomials of degree 11 or less with equals B −1 A−1 which does not usually equal
coefficients in R. True. They’re both dimension A−1 B −1 .
12, so they’re isomorphic.
e. (AB)2 = A2 B 2 . Not identity. (AB)2 equals
e. If A is an invertible matrix, then AB = AC ABAB, and since BA does not usually equal AB,
implies that B = C. True. If you multiply the ABAB does not usually equal A2 B 2 .
first equation on the left by A−1 , you get A−1 AB =
A−1 AC, which simplifies to the second equation. f. (A2 + I)(A2 − I) = A4 − I. Identity. (A2 +
I)(A2 − I) = A4 − A2 + A2 − I = A4 − I.
f. If the kernel of a transformation is 0, then the
transformation is one-to-one. True. This is one of
the reasons kernels are so important.
g. A linear transformation is invertible if and
only if it is one-to-one and onto. True. If it’s one-
to-one and onto, then there’s an inverse function,
and we showed that it’s linear.

4. [16] Recall that two square matrices A and B

are said to be similar or conjugate when there is an
invertible matrix Q such that B = Q−1 AQ. We’ll
denote similar matrices A ∼ B. Prove that this
relation is transitive, that is, show that if one ma-
trix is similar to another, and the second is similar
to the third, then the first is similar to the third.
Symbolically, that says A ∼ B and B ∼ C imply
A ∼ C.
Here’s one proof. Since A ∼ B, there’s an in-
vertible matrix Q that conjugates A to B, that is,
B = Q−1 AQ. Since B ∼ C, there’s an invert-
ible matrix that conjugates B to C, but that could
be a different matrix than Q, so call it P . Then
C = P −1 BP . Therefore,
C = P −1 BP = P −1 Q−1 AQP = (QP )−1 A(QP )
which shows that C is a conjugate of A. q.e.d.