Math 215 HW #8 Solutions

1. Problem 4.2.4. By applying row operations to produce an upper triangular U, compute

det

1 2 2 0
2 3 4 1
1 2 0 2
0 2 5 3

and det

2 1 0 0
1 2 1 0
0 1 2 1
0 0 1 2

.
Answer: Focusing on the rst matrix, we can subtract twice row 1 from row 2 and add row
1 to row 3 to get

1 2 2 0
0 1 0 1
0 0 2 2
0 2 5 3

.
Next, add twice row 2 to row 4:

1 2 2 0
0 1 0 1
0 0 2 2
0 0 5 5

.
Finally, add 5/2 times row 3 to row 4:

1 2 2 0
0 1 0 1
0 0 2 2
0 0 0 10

.
Since none of the above row operations changed the determinant and since the determinant
of a triangular matrix is the product of the diagonal entries, we see that
det

1 2 2 0
2 3 4 1
1 2 0 2
0 2 5 3

= (1)(1)(2)(10) = 20.
Turning to the second matrix, we can rst add half of row 1 to row 2:

2 1 0 0
0 3/2 1 0
0 1 2 1
0 0 1 2

.
Next, add 2/3 of row 2 to row 3:

2 1 0 0
0 3/2 1 0
0 0 4/3 1
0 0 1 2

.
1
Finally, add 3/4 of row 3 to row 4:

2 1 0 0
0 3/2 1 0
0 0 4/3 1
0 0 0 5/4

.
Therefore, since the row operations didnt change the determinant and since the determinant
of a triangular matrix is the product of the diagonal entries,
det

2 1 0 0
1 2 1 0
0 1 2 1
0 0 1 2

= (2)(3/2)(4/3)(5/4) =
5!
4!
= 5.
Note: This second matrix is the same one that came to our attention in Section 1.7 and
HW #3, Problem 9.
2. Problem 4.2.6. For each n, how many exchanges will put (row n, row n 1, . . ., row 1) into
the normal order (row 1, . . ., row n 1, row n)? Find det P for the n by n permutation with
1s on the reverse diagonal.
Answer: Suppose n = 2m is even. Then the following sequence of numbers gives the original
ordering of the rows:
2m, 2m1, . . . , m+ 1, m, . . . , 2, 1.
Exchanging 2m and 1, and then 2m1 and 2, . . ., and then m+ 1 and m yields the correct
ordering of rows:
1, 2, . . . , m, m+ 1, . . . , 2m1, 2m.
Clearly, we performed m = n/2 row exchanges in the above procedure. Thus, for even values
of n, we need to perform n/2 row exchanges.
On the other hand, suppose n = 2m1 is odd. Then the original ordering of the rows is
2m1, 2m2, . . . , m+ 1, m, m1, . . . , 2, 1.
We exchange 2m1 and 1, and then 2m2 and 2, . . ., and then m+1 and m1. Since m
is already in the correct spot, this gives the correct ordering of rows
1, 2, . . . , m1, m, m+ 1, . . . , 2m2, 2m1.
Clearly, we performed m 1 =
n1
2
row exchanges. Thus, for odd values of n, we need to
perform
n1
2
row exchanges.
If P is the permutation matrix with 1s on the reverse diagonal, then the rows of P are simply
the rows of the identity matrix in precisely the reverse order. Thus, the above reasoning tells
us how many row exchanges will transform P into I. Since the determinant of the identity
matrix is 1 and since performing a row exchange reverses the sign of the determinant, we
have that
det P = (1)
number of row exchanges
det I = (1)
number of row exchanges
.
2
Therefore,
det P =

(1)
n/2
if n is even
(1)
n1
2
if n is odd
=

1 if
n
4
has remainder 0 or 1
1 if
n
4
has remainder 2 or 3
.
3. Problem 4.2.8. Show how rule 6 (det = 0 if a row is zero) comes directly from rules 2 and 3.
Answer: Suppose A is an nn matrix such that the ith row of A is equal to zero. Let B be
the matrix which comes from exchanging the rst row and the ith row of A. Then, by rule 2,
det B = det A.
Now, the matrix B has all zeros in the rst row. Therefore, by rule 3,
det B =

0 0 0
b
21
b
22
b
2n
.
.
.
.
.
.
.
.
.
b
n1
b
n2
b
nn

0 1 0 1 0 1
b
21
b
22
b
2n
.
.
.
.
.
.
.
.
.
b
n1
b
n2
b
nn

= 0

1 1 1
b
21
b
22
b
2n
.
.
.
.
.
.
.
.
.
b
n1
b
n2
b
nn

= 0.
Since det B = 0 and since det A = det B, we see that
det A = det B = 0 = 0,
which is rule 6.
4. Problem 4.2.10. If Q is an orthogonal matrix, so that Q
T
Q = I, prove that det Q equals +1
or 1. What kind of box is formed from the rows (or columns) of Q?
Answer: By rule 10, we know that det(Q
T
) = det Q. Therefore, using rules 1 and 9,
1 = det I = det(Q
T
Q) = det(Q
T
) det Q = (det Q)
2
.
Hence,
det Q =

1 = 1.
We see that the columns of Q form a box of volume 1. In fact, they form a cubical box.
5. Problem 4.2.14. True or false, with reason if true and counterexample if false.
(a) If A and B are identical except that b
11
= 2a
11
, then det B = 2 det A.
Answer: False. Suppose
A =

1 1
1 1

B =

2 1
1 1

.
Then det A = 0 and det B = 2 1 = 1 = 2 det A.
(b) The determinant is the product of the pivots.
Answer: False. Let
A =

0 1
1 0

.
Then det A = 0 1 = 1, but the two pivots are 1 and 1, so the product of the pivots is
1. (The issue here is that we have to do a row exchange before we try elimination and
the row exchange changes the sign of the determinant)
3
(c) If A is invertible and B is singular, then A+B is invertible.
Answer: False. Let
A =

1 0
0 1

B =

1 0
0 0

.
Then A, being the identity matrix, is invertible, while B, since it has a row of all zeros,
is denitely singular. However,
A+B =

0 0
0 1

is singular since it has a zero row.

(d) If A is invertible and B is singular, then AB is singular.
Answer: True. Since B is singular, det B = 0. Therefore,
det(AB) = det Adet B = det A 0 = 0.
Since det(AB) = 0 only if AB is singular, we can conclude that AB is singular.
(e) The determinant of AB BA is zero.
Answer: False. Let
A =

0 1
2 0

B =

0 3
5 0

.
Then
AB =

0 1
2 0

0 3
5 0

5 0
0 6

and
BA =

0 3
5 0

0 1
2 0

6 0
0 5

.
Therefore,
AB BA =

1 0
0 1

,
which has determinant equal to 1.
6. Problem 4.2.26. If a
ij
is i times j, show that det A = 0. (Exception when A = [1]).
Proof. Notice that the rst row of A is
[1 2 3 4 n]
and the second row of A is
[2 4 6 8 2n].
Thus, the rst two rows of A are linearly dependent, meaning that A is singular since elim-
ination will produce a row of all zeros in the second row. Thus, the determinant of A must
be zero. (In fact, every row is a multiple of the rst row, so A is about as far as a non-zero
matrix can be from being non-singular).
4
7. Problem 4.3.6. Suppose A
n
is the n by n tridiagonal matrix with 1s on the three diagonals:
A
1
= [1], A
2
=

1 1
1 1

, A
3
=

1 1 0
1 1 1
0 1 1

, . . .
Let D
n
be the determinant of A
n
; we want to nd it.
(a) Expand in cofactors along the rst row to show that D
n
= D
n1
D
n2
.
Proof. We want to nd the determinant of
A
n
=

1 1 0 0 0
1 1 1 0 0
0 1 1 1 0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0 0 0 0 1

.
Doing a cofactor expansion along the rst row, D
n
will be equal to 1 times the deter-
minant of the matrix given by deleting the rst row and rst column minus 1 times the
determinant of the matrix given by deleting the rst row and second column.
Deleting the rst row and rst column of A
n
just leaves a copy of A
n1
, the determinant
of which is D
n1
. Thus,
D
n
= 1 D
n1
1 det(matrix left when deleting rst row and second column). (1)
Deleting the rst row and second column yields the matrix

1 1 0 0
0 1 1 0
.
.
.
.
.
.
.
.
.
.
.
.
0 0 0 1

. (2)
Notice that if we delete the rst row and rst column of this matrix, were left with
a copy of A
n2
(the determinant of which is D
n2
), whereas when we delete the rst
row and second column we get a matrix with all zeros in the rst column (which must
have determinant zero). Thus, the determinant of the matrix from (2) is, using cofactor
expansion, equal to
1 D
n2
1 0.
Therefore, combining this with (1), we see that
D
n
= 1 D
n1
1 (1 D
n2
1 0)
or, equivalently,
D
n
= D
n1
D
n2
.
5
(b) Starting from D
1
= 1 and D
2
= 0, nd D
3
, D
4
, . . . , D
8
. By noticing how these numbers
cycle around (with what period?) nd D
1000
.
Answer: Since D
1
= 1 and D
2
= 0, we have, using the result from part (a), that
D
3
= D
2
D
1
= 0 1 = 1
D
4
= D
3
D
2
= 1 0 = 1
D
5
= D
4
D
3
= 1 (1) = 0
D
6
= D
5
D
4
= 0 (1) = 1
D
7
= D
6
D
5
= 1 0 = 1
D
8
= D
7
D
6
= 1 1 = 0
.
.
.
Since each term depends only on the two preceding terms and since D
8
= D
2
and
D
7
= D
1
, the above pattern will repeat indenitely. Thus, the Ds have a period of
7 1 = 6, so D
1+6m
= D
1
for each m and, more generally, D
k+6m
= D
k
for any m,
where k {1, 2, 3, 4, 5, 6}. Therefore,
D
1000
= D
4+6166
= D
4
= 1.
8. Problem 4.3.8. Compute the determinants of A
2
, A
3
, A
4
. Can you predict A
n
?
A
2
=

0 1
1 0

A
3
=

0 1 1
1 0 1
1 1 0

A
4
=

0 1 1 1
1 0 1 1
1 1 0 1
1 1 1 0

.
Use row operations to produce zeros, or use cofactors of row 1.
Answer: Using the formula for determinants of 2 2 matrices, we see that
det(A
2
) = 0 0 1 1 = 1.
Then, expanding in cofactors along the rst row,
det(A
3
) = 0

0 1
1 0

1 1
1 0

+ 1

1 0
1 1

= 0 1(1) + 1(1)
= 2.
Again, doing a cofactor expansion along the rst row,
det(A
4
) = 0

0 1 1
1 0 1
1 1 0

1 1 1
1 0 1
1 1 0

+ 1

1 0 1
1 1 1
1 1 0

1 0 1
1 1 0
1 1 1

= 0 1

1 1 1
1 0 1
1 1 0

+ 1 (1)

1 0 1
1 1 1
1 1 0

1 (1)
2

1 0 1
1 1 0
1 1 1

(3)
6
using Property 2 of the determinant (which says that exchanging rows changes the sign of
the determinant). Now,

1 1 1
1 0 1
1 1 0

= 1

0 1
1 0

1 1
1 0

+ 1

1 0
1 1

= 1(1) 1(1) + 1(1) = 1,

so, plugging this into (3), we see that
det(A
4
) = 0 1(1) + 1(1)(1) 1(1)(1) = 3.
In general, it will turn out that
det(A
n
) = (1)
n1
(n 1).
9. Problem 4.3.14. Compute the determinants of A, B, C. Are their columns independent?
A =

1 1 0
1 0 1
0 1 1

B =

1 2 3
4 5 6
7 8 9

C =

A 0
0 B

.
Answer: First, compute the determinant of A using cofactors:
det A = 1

0 1
1 1

1 1
0 1

+ 0

1 0
0 1

= 1(1) 1(1) + 0(1)

= 2.
Since det A = 0, the matrix A is invertible and thus the columns of A are necessarily linearly
independent.
Next, compute the determinant of B using cofactors:
det B = 1

5 6
8 9

4 6
7 9

+ 3

4 5
7 8

= 1(5 9 6 8) 2(4 9 6 7) + 3(4 8 5 7)

= 3 + 12 9
= 0.
Thus, since det B = 0, the matrix B is not invertible and so its columns are not linearly
independent.
Turning attention to the matrix C, note that, since the columns of B are linearly dependent,
the last three columns of C must also be linearly dependent, meaning that det C = 0.
10. Problem 4.3.28. The n by n determinant C
n
has 1s above and below the main diagonal:
C
1
= |0| C
2
=

0 1
1 0

C
3
=

0 1 0
1 0 1
0 1 0

C
4
=

0 1 0 0
1 0 1 0
0 1 0 1
0 0 1 0

.
7
(a) What are the determinants C
1
, C
2
, C
3
, C
4
?
Answer: Clearly, det C
1
= |0| = 0. Next,
det C
2
=

0 1
1 0

= 0 0 1 1 = 1.
Now, expanding in cofactors,
det C
3
= 0

0 1
1 0

1 1
0 0

+ 0

1 0
0 1

= 0(1) 1(0) + 0(1)

= 0.
Finally, we also determine det C
4
by expanding in cofactors:
det C
4
= 0

0 1 0
1 0 1
0 1 0

1 1 0
0 0 1
0 1 0

+ 0

1 0 0
0 1 1
0 0 0

1 0 1
0 1 0
0 0 1

= 1

1 1 0
0 0 1
0 1 0

= 1

0 1
1 0

0 1
0 0

+ 0

0 0
0 1

= 1(1 det C
2
)
= 1.
(b) By cofactors nd the relation between C
n
and C
n1
and C
n2
. Find C
10
.
Answer: Just as in the n = 4 case, doing a cofactor expansion along the rst row yields
only one non-zero term, namely
det C
n
= 1

1 1 0 0 0
0 1 1 0 0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0 0 0 1 0

Deleting the rst row and second column yields a matrix with all zeros in the rst
column, which necessarily has determinant zero. Therefore, using a cofactor expansion,
the above is equal to
det C
n
= 1 (1 det C
n2
1 0) = det C
n2
.
Thus, we have that det C
n
= det C
n2
. Hence,
det C
10
= det C
8
= det C
6
= det C
4
= 1.
11. Let the numbers S
n
be the determinants dened in Problem 4.3.31.
8
(a) For any n > 2 prove that S
n
= 3S
n1
S
n2
.
Proof. We can compute S
n
using a cofactor expansion:
S
n
=

3 1 0 0
1 3 1 0
0 1 3 0
.
.
.
.
.
.
.
.
.
.
.
.
0 0 0 3

= 3

3 1 0
1 3 0
.
.
.
.
.
.
.
.
.
0 0 3

1 1 0
0 3 0
.
.
.
.
.
.
.
.
.
0 0 3

= 3S
n1
1

1 1 0
0 3 0
.
.
.
.
.
.
.
.
.
0 0 3

.
Doing a cofactor expansion of this new determinant gives 1 S
n2
plus 1 times the
determinant of a matrix with all zeros in the rst column. Thus, the second term in the
above expression is just 1 S
n2
, so we can conclude that
S
n
= 3S
n1
S
n2
.
(b) For any k let F
k
denote the kth Fibonacci number (recall that the Fibonacci sequence
1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, . . . is dened by F
k
= F
k1
+ F
k2
). Prove that
F
2n+2
= 3F
2n
F
2n2
.
Proof. By denition of the Fibonacci sequence, we know that
F
2n+2
= F
2n+1
+F
2n
F
2n+1
= F
2n
+F
2n1
F
2n
= F
2n1
+F
2n2
.
From the third line, we have that F
2n1
= F
2n
F
2n2
. Therefore, substituting the
second line into the rst and using this expression for F
2n1
, we have that
F
2n+2
= F
2n+1
+F
2n
= (F
2n
+F
2n1
) +F
2n
= (F
2n
+ (F
2n
F
2n2
)) +F
2n
= 3F
2n
F
2n2
.
9
(c) Show that S
n
= F
2n+2
for each n.
Proof. I will prove this using the principle of mathematical induction. Let P
k
be the
statement that S
k
= F
2k+2
.
The base case of induction is to prove that P
1
and P
2
are true; i.e. that S
1
= F
4
and
S
2
= F
6
. However, both are clearly true, as
S
1
= 3 = F
4
and S
2
=

3 1
1 3

= 8 = F
6
.
For the inductive step, we want to show that P
k
and P
k1
being true implies P
k+1
is
true for any k. To see this, suppose P
k
and P
k1
are true, meaning that
S
k
= F
2k+2
and S
k1
= F
2k
.
Then, using part (a),
S
k+1
= 3S
k
S
k1
= 3F
2k+2
F
2k
.
However, by part (b), the right-hand side of the equation is equal to F
2k+4
, so we see
that S
k+1
= F
2k+4
, which is to say that P
k+1
is true.
Therefore, since weve shown that P
1
and P
2
are true and weve shown that P
k
and P
k1
being true implies P
k+1
is true, so, by induction, we can conclude that P
n
is true for all
n. In other words,
S
n
= F
2n+2
for all n.
12. (Bonus Problem) Problem 3.5.12. Compute F
8
c by the three steps of the Fast Fourier
Transform if c = (1, 0, 1, 0, 1, 0, 1, 0). Repeat the computation with c = (0, 1, 0, 1, 0, 1, 0, 1).
Answer: Note, rst of all, that
c

1
1
1
1

and

0
0
0
0

.
Hence,
y

= F
4
c

1 1 1 1
1 i 1 i
1 1 1 1
1 i 1 i

1
1
1
1

4
0
0
0

and
y

= F
4
c

= F
4

0 =

0.
10
Therefore,
y
1
= y

1
+w
8
y

1
= 4 + 0 = 4
y
2
= y

2
+w
2
8
y

2
= 0 + 0 = 0
y
3
= y

3
+w
3
8
y

3
= 0 + 0 = 0
y
4
= y

4
+w
4
8
y

4
= 0 + 0 = 0
y
5
= y

1
w
5
8
y

1
= 4 0 = 4
y
6
= y

2
w
6
8
y

2
= 0 0 = 0
y
3
= y

3
w
7
8
y

3
= 0 0 = 0
y
4
= y

4
w
8
8
y

4
= 0 0 = 0,
where w
8
is an eighth root of 1. Therefore,
y =

4
0
0
0
4
0
0
0

.
Turning to c = (0, 1, 0, 1, 0, 1, 0, 1), we see that
c

0
0
0
0

and

1
1
1
1

.
Hence,
y

= F
4
c

= F
4

0 =

0
and
y

= F
4
c

1 1 1 1
1 i 1 i
1 1 1 1
1 i 1 i

1
1
1
1

4
0
0
0

.
11
Therefore, since w
8
=
1

2
+i
1

2
, we see that
y
1
= y

1
+w
8
y

1
= 0 +

2
+i
1

4 = 2

2 + 2

2i.
y
2
= y

2
+w
2
8
y

2
= 0 + 0 = 0
y
3
= y

3
+w
3
8
y

3
= 0 + 0 = 0
y
4
= y

4
+w
4
8
y

4
= 0 + 0 = 0
y
5
= y

1
w
5
8
y

1
= 0

2
i
1

4 = 2

2 + 2

2i
y
6
= y

2
w
6
8
y

2
= 0 0 = 0
y
3
= y

3
w
7
8
y

3
= 0 0 = 0
y
4
= y

4
w
8
8
y

4
= 0 0 = 0.
Then
y =

2 + 2

2i
0
0
0
2

2 + 2

2i
0
0
0

.
12