APP. A] VECTORS AND MATRICES 425
A.11. Find the transpose of each matrix:
A=
[
1 − 23
78 − 9
]
; B=
⎡
⎣
123
245
356
⎤
⎦; C=[ 1 ,− 3 , 5 ,− 7 ]; D=
⎡
⎣
2
− 4
6
⎤
⎦
Rewrite the rows of each matrix as columns to obtain the transposes of the matrices:
AT=
⎡
⎣
17
− 28
3 − 9
⎤
⎦,BT=
⎡
⎣
123
245
356
⎤
⎦,CT=
⎡
⎢⎢
⎣
1
− 3
5
− 7
⎤
⎥⎥
⎦,D
T=[ 2 ,− 4 , 6 ]
(Note thatBT=B; such a matrix is said to besymmetric. Note also that the transpose of the row vectorCis a column
vector, and the transpose of the column vectorDis a row vector.)
A.12. Prove Theorem A.2(i):A(BC)=A(BC).
LetA=[aij],B=[bjk], andC=[ckl]. Furthermore, letAB=S=[sik]andBC=T=[tjl].
Then
sik=ai 1 b 1 k+ai 2 b 2 k+···+aimbmk=
∑m
j= 1
aijbjk
tjl=bj 1 c 1 i+bj 2 c 2 i+···+bjncnl=
∑n
k=l
bjkckl
Now, multiplyingSbyC, i.e., (AB)byC, the element in theith row andlth column of the matrix(AB)Cis
si 1 c 1 l+si 2 c 2 l+···+sincnl=
∑n
k= 1
sikckl=
∑n
k= 1
∑m
j= 1
(aijbjk)ckl
On the other hand, multiplyingAbyT, i.e.,AbyBC, the element in theith row andlth column of the matrix
A(BC)is
ai 1 t 1 l+ai 2 t 2 l+···+aimtml=
∑m
j= 1
aijtjl=
∑m
k= 1
∑n
j= 1
aij(bjkckl)
Since the above sums are equal, the theorem is proven.
SQUARE MATRICES, DETERMINANTS, INVERSES
A.13. Find the diagonal and trace of each matrix:
(a)A=
⎡
⎣
136
2 − 58
4 − 27
⎤
⎦;(b)B=
[
t− 23
− 4 t+ 5
]
;(c)C=
[
12 − 3
4 − 56
]
.
(a) The diagonal consists of the elementsa 11 ,a 22 ,a 33 , that is, the scalars 1,− 5 ,7. The trace is the sum of the diagonal
elements; hence tr(A)= 1 − 5 + 7 =3.
(b) The diagonal consists of the pair{t− 2 ,t+ 5 }. Thus tr(B)=t− 2 +t+ 5 = 2 t+3.
(c) The diagonal and trace are defined only for square matrices.
A.14. Let A =
[
12
4 − 3
]
. Find: (a)A^2 ;(b)A^3 ;(c)f (A) where f(x) = 2 x^3 − 4 x +5;
(d) g(A)whereg(x)=x^2 + 2 x−11.
(a) A^2 =AA=
[
12
4 − 3
][
12
4 − 3
]
=
[
1 + 82 − 6
4 −12 8+ 9
]
=
[
9 − 4
− 817
]
.
(b) A^3 =AA^2 =
[
12
4 − 3
][
9 − 4
− 817
]
=
[
9 − 16 − 4 + 34
36 + 24 − 16 − 51
]
=
[
− 730
60 − 67
]
.