解答僅供參考，請同學務必詳細描述計算 (or 證明) 過程

作業題目是 8、9、11、12、13、14、15、17、22、23、24、25、29、30、33、
35、39、40，其他題的解答供同學練習
Exercise 3.1.8
Since B has three columns and B T has three rows, the product makes sense,
and  
[ ] 4 0 [ ]
4 −2 1 −2 2 = 21 −1
BB T =
0 2 3 −1 13
1 3

Exercise 3.1.9
Since A has two columns andF has two rows, AF makes sense; since A has two
rows and F has one column, the result is a 2 × 1 matrix. So this matrix has
two rows, and E has two columns, so we can multiply by E on the left. Thus
this product makes sense. Then
[ ][ ] [ ]
3 0 −1 −3
AF = =
−1 5 2 11
and then [ ]
[ ] −3
E(AF ) = 4 2 = [10]
11

Exercise 3.1.10
F has shape 2 × 1. D has shape 2 × 2, so DF has shape 2 × 1. Since the
number of columns(1) of F does not equal the number of rows(2) of DF, the
product does not make sense.
Exercise 3.1.11
SinceF has one column and E has one row, the product makes sense, and
[ ] [ ]
−1 [ ] −4 −2
FE = 4 2 =
2 8 4

Exercise 3.1.12
Since E has two columns and F has one row, the product makes sense, and
[ ]
[ ] −1 [ ]
EF = 4 2 = 0
2

1
(Note that FE from Exercise 11 and EF are not the same matrix. In fact,
they don’t even have the same shape! In general, matrix multiplication is not
commutative.)
Exercise 3.1.13
Since B has two rows, B T has two columns; since C has two columns, C T has
two rows. Thus B T C T makes sense, and the result is a 3 × 3 matrix (since
B T has three rows and C T has three columns). Next, CB is defined (3 × 2
multiplied by 2 × 3) and also yields a 3 × 3 matrix, so its transpose is also a 3
× 3 matrix. Thus B T C T −(BC)T is defined.
   
4 0 [ ] 4 12 20
1 3 5
B T C T = −2 2 = 2 2 2 
2 4 6
1 3 7 15 23
   
1 2 [ ] 4 2 7
4 −2 1
CB = 3 4 = 12 2 15
0 2 3
5 6 20 2 23
 
4 12 20
(CB)T = 2 2 2 
7 15 23

Hence B T C T −(CB)T is the zero matrix.

2
Exercise 3.1.14
Since A and D are both 2 × 2 matrices, the products in both orders are defined,
and each results in a 2 × 2 matrix, so their difference is defined as well.
[ ][ ] [ ]
0 −3 3 0 3 −15
DA = =
−2 1 −1 5 −7 5
[ ][ ] [ ]
3 0 0 −3 0 −9
AD = =
−1 5 −2 1 −10 8
[ ]
3 −6
AD − DA =
3 −3

Exercise 3.1.15
Since A is 2 × 2, A2 is defined and is also a 2 × 2 matrix, so we can multiply
by A again to get A3

[ ][ ] [ ]
2 3 0 3 0 9 0
A = =
−1 5 −1 5 −8 25
[ ][ ] [ ]
3 3 0 9 0 27 0
A = =
−1 5 −8 25 −49 125

Exercise 3.1.17
[ ]
a c
Suppose that A = . Then if A2 = 0,
b d
[ ][ ] [ 2 ] [ ]
a c a c a + bc ab + cd 0 0
A2 = = =
b d b d ac + cd bc + d2 0 0

If c = 0, so that the upper right-hand entry is zero, then[ we]must also have
1 0
a = 0 and d = 0 so that the diagonal entries are zero. So is one solution
0 0
Alternatively, if c ̸= 0, then ac = −cd and thus a = −d. This will make the
lower left entry zero as well, and the two diagonal entries will both be equal to
a2 + bc. So any values of a, b, and c ̸= 0 with a2 = −bc will work. For example,
a = b = 1, c = −1, d = −1, then
[ ][ ] [ ]
1 1 1 1 0 0
=
−1 −1 −1 −1 0 0

3
Exercise 3.1.22

    
−1 0 2 x1 1
 1 −1 0 x2  = −2
0 1 1 x3 −1
, so that Ax = b , where
     
−1 0 2 x1 1
    
A = 1 −1 0 , x = x2 , b = −2
0 1 1 x3 −1

Exercise 3.1.23
The column vectors of B are
     
2 3 0
  
b1 = 1 , b2 = −1 , b3 = 1
 
−1 6 4
so the matrix-column representation is
       
1 0 −2 4
    
Ab1 = 2 −3 + 1 1 − 1 1 = −6  
−2 0 −1 5
       
1 0 −2 −9
Ab2 = 3 −3 − 1 1 + 6  1  = −4
−2 0 −1 0
       
1 0 −2 −8
Ab3 = 0 −3 + 1 1 + 4 1 = 5 
      
−2 0 −1 −4
Thus  
4 −9 −8
AB = −6 −4 5 
5 0 −4

Exercise 3.1.24
The row vectors of A are
[ ] [ ] [ ]
A1 = 1 0 −2 , A2 = −3 1 −2 , A3 = 2 0 −1

4
so the row-matrix representation is
[ ] [ ] [ ] [ ]
A1 B = 1 2 3 0 + 0 1 −1 1 − 2 −1 6 4 = 4 −9 −8
[ ] [ ] [ ] [ ]
A2 B = −3 2 3 0 + 1 1 −1 1 − 1 −1 6 4 = −6 −4 −5
[ ] [ ] [ ] [ ]
A3 B = 2 2 3 0 + 0 1 −1 1 − 1 −1 6 4 = 5 0 −4

Thus  
4 −9 −8
AB = −6 −4 −5
5 0 −4

Exercise 3.1.25
The outer product expansion is
     
1 [ ] 0 [ ] −2 [ ]
a1 B1 + a2 B2 + a3 B3 = −3 2 3 0 + 1 1 −1 1 +  1  −1 6 4
2 0 −1
     
2 3 0 0 0 0 2 −12 −8

= −6 −9 0 + 1 −1 1 + −1 6 4
4 6 0 0 0 0 1 −6 −4
 
4 −9 −8
= −6 −4 5
5 0 −4

Exercise 3.1.29
[ ]
Assume that the columns of B = b1 b2 . . . bn are linearly dependent. Then
there exists a solution to x1 b1 + x2 b2 + · · · + xn bn = 0 with at least one xi ̸= 0.
Now, we have
[ ] [ ]
AB = A b1 b2 . . . bn = Ab1 Ab2 . . . Abn

so that A(x1 b1 + x2 b2 + · · · + xn bn ) = x1 (Ab1 ) + x2 (Ab2 ) + · · · + xn (Abn ) = 0,

5
Exercise 3.1.30
Let    
a1 a1 B
a  a B 
 2  2 
A =  ..  , so that AB =  .. 
.  . 
an an B

Since the rows of A are linearly dependent, there are xi , not all zero, such
that x1 a1 + x2 a2 + · · · + xn an = 0. But then (x1 a1 + x2 a2 + · · · + xn an )B =
x1 (a1 B) + x2 (a2 B) + ··· + xn (an B) = 0, so that the rows of AB are linearly
dependent.
Exercise 3.1.33
We have the block structure
[ ][ ] [ ]
A11 A12 B11 B12 A11 B11 + A12 B21 A11 B12 + A12 B22
AB = =
A21 A22 B21 B22 A21 B11 + A22 B21 A21 B12 + A22 B22
 
[ ] 1 2 2 0
A11 A11 B12 + B22 3 4 5 3 
= =1 0 1 2 

I2 B12 + A22 B22
0 1 0 −1

Exercise 3.1.34
We have the block structure
[ ][ ] [ ]
A11 A12 B11 B12 A11 B11 + A12 B21 A11 B12 + A12 B22
AB = =
A21 A22 B21 B22 A21 B11 + A22 B21 A21 B12 + A22 B22
Now, A11 is the identity matrix I3 , and A21 is the zero matrix. Further, B22 is
a 1 × 1 matrix, as is A22 , so they both behave just like scalar multiplication.
Thus this product reduces to
[ ] [ ]
I3 B11 + A12 B21 I3 B12 − A12 B11 + A12 B21 B12 − A12
AB = =
4B21 −4 4B21 −4
Compute the matrices in the first row:
       
1 [ ] 1 2 3 1 1 1 2 3 4
B11 + A12 B21 = B11 + 2 1 1 1 = 0 1 4 + 2 2 2 = 2 3 6
3 0 0 1 3 3 3 3 3 4
     
1 1 0
B12 − A12 = 1 − 2 = −1
1 3 2

6
So  
2 3 4 0
2 3 6 −1
AB = 
3

3 4 −2
4 4 4 −4

Exercise 3.1.35
(a) Computing the powers of A as required,we get
[ ] [ ] [ ]
−1 1 −1 0 0 −1
A2 = , A3 = , A4 = ,
−1 0 0 −1 1 −1
[ ] [ ] [ ]
1 −1 1 0 0 1
A5 = , A6 = = I2 , A 7 = = A,
1 0 0 1 −1 1
Note that A6 = I2 and A7 = A.
(b) SinceA6 = I2 ,we know that A6k = (A6 )k = I2k = I2 for any positive integer
value of k. Since 2015 = 2010 + 5 = 6 · 370 + 5,we have
[ ]
1 −1
A2015 = A6·370+5 = A6·370 A5 = I2 A5 = A5 =
1 0
Exercise 3.1.38
(a)
[ ][ ] [ 2 ]
cos θ − sin θ cos θ − sin θ cos θ − sin2 θ −2 cos θ sin θ
A2 = =
sin θ cos θ sin θ cos θ 2 cos θ sin θ cos2 θ − sin2 θ

Since cos2 θ − sin2 θ = cos 2θ and 2 cos θ sin θ = sin 2θ, so we get
[ ]
2 cos 2θ − sin 2θ
A =
sin 2θ cos 2θ

(b)
We have already proven the cases n = 1 and n = 2, which serve as the basis for
the induction. Now assume that the given formula holds for k. Then

inductive [ ][ ]
k hypothesis cos θ − sin θ cos kθ − sin kθ
Ak+1 = A1 A =
sin θ cos θ sin kθ cos kθ
[ ]
cos θ cos kθ − sin θ sin kθ −(cos θ sin kθ + sin θ cos kθ
=
sin θ cos kθ + cos θ sin kθ cos θ cos kθ − sin θsinkθ

7
But cos θ cos kθ − sin θ sin kθ = cos θ + kθ = cos (k + 1)θ, and sin θ + cos kθ +
cos θ sin kθ = sin (θ + kθ) = sin (k + 1)θ, so this matrix becomes
[ ]
cos (k + 1)θ − sin (k + 1)θ
Ak+1 =
sin (k + 1)θ cos (k + 1)θ
So by induction, the formula holds for all k ≥ 1.
Exercise 3.1.39
(a) aij = (−1)i+j means that if i + j is even, then the corresponding entry is 1;
otherwise, it is -1  
1 −1 1 −1
−1 1 −1 1 
A=  1 −1 1 −1


−1 1 −1 1

(b) Each entry is its row number minus its column number:
 
0 −1 −2 −3
1 0 −1 −2
A= 2 1
.
0 −1
3 2 1 0

(c)
   
(1 − 1)1 (1 − 1)2 (1 − 1)3 (1 − 1)4 0 0 0 0
(2 − 1)1 (2 − 1)2 (2 − 1)3 (2 − 1)  
4
1
A= 1 1 1 .
(3 − 1)1 4 = 
(3 − 1)2 (3 − 1)3 (3 − 1) 2 4 8 16
(4 − 1)1 (4 − 1)2 (4 − 1)3 (4 − 1) 4
3 9 27 81

(d)  
(1+1−1)π (1+2−1)π
sin 4 sin 4 sin (1+3−1)π
4 sin (1+4−1)π
4
 (2+1−1)π (2+2−1)π (2+3−1)π (2+4−1)π 
sin sin sin sin 
A= 4 4 4
(3+4−1)π 
4
sin (3+1−1)π
4 sin(3+2−1)π
4 sin (3+3−1)π
4 sin 4

(4+1−1)π (4+2−1)π
sin 4 sin 4 sin (4+3−1)π
4 sin (4+4−1)π
4
 
sin π4 sin π2
sin 3π 4 sin π
 sin π
sin 3π sin π sin 5π 
=  3π 2 4 4 
sin sin π sin 5π 3π 
4 4 sin 2
sin π sin 5π 3π
4 sin 2 sin 4
7π
 √2 √
2

2 √
1 2 0
√
 
 √1 2
0√ − 22 
= 2 2 .
2 0√ − 22 −1 
√
0 − 22 −1 − 22

8
Exercise 3.1.40
(a)
This matrix has zeros whenever the row number exceeds the column number,
and the nonzero entries are the sum of the row and column numbers:
 
2 3 4 5 6 7
0 4 5 6 7 8 
 
0 0 6 7 8 9 
A= 0 0 0 8 9 10

 
0 0 0 0 10 11
0 0 0 0 0 12

(b)
This matrix has 1’s on the main diagonal and one diagonal away from the main
diagonal, and zeros elsewhere:
 
1 1 0 0 0 0
1 1 1 0 0 0 
 
0 1 1 1 0 0 
A= 0 0 1 1 1 0 

 
0 0 0 1 1 1 
0 0 0 0 1 1

(c)
This matrix is zero except where the row number plus the column number is 6,
7, or 8:  
0 0 0 0 1 1
0 0 0 1 1 1 
 
0 0 1 1 1 0 
A= 0 1 1 1 0 0 

 
1 1 1 0 0 0 
1 1 0 0 0 0