CHAPTER 4

|Vectors and Matrices

VECTORS IN R" operations:

4.1 following vector + b,,..., a, + b,)
u +v= (a, b,, a,
+
section considers the Vector Addition: (ka,, , ka,)
ka,,
This Multiplication: ku =
Scalar =(-a,, -a,, , -a,)
Negation: -u =(-)u vectors in R" and kkis a real number
b,..., b,) are (scalar).
a.) andv=(b,,
where u= (a,, a2, . be vectors in R Which of the vectors,if any, are
= (3, 5, 1), u, =(1,. 5, 3), u, =(3, 5, 1)
u,
4.1 Let u, =(1, 3, 5),
equal? u, = u,.
componentwise equal; hence only
Only u, and u, are
n
-). Find: (q) utu, (6) v+ w, (c) -3,
4.2
Iet u =(2. -7, 1), v=(-3, 0, 4), w=(0, 3,
components:
(a) Add corresponding
u+y= (2, -7, 1) + (-3, 0, 4) =(2-3, -7+0, 1 +4) =(-1, -7, 5)
components:
(6) Add corresponding
y+w=(-3, 0, 4) +(0, 5, -8) =(-3+0,0+5, 4- 8) =(-3, 5, -4)
by the scalar -3:
(c) Multiply each component of u
-3u =-3(2, -7, 1) =(-6, 21, -3)
(d) Multiply each component of w by -1, i.e., change the sign of each component
-w=-(0, 5, -8) =(0, -5, 8)

4.3 Let u. u, andwbe the vectors of Problem 4.2. Find: (a) 3u - 4v and (b)2u + 3u-5w.
| First perform the scalar multiplication and then the vector addition.
(a) 3u - 4u = 3(2,-7, 1) 4-3, 0,4) =(6, -21, 3) + (12, 0, -16) = (18, -21, -13)
(b) 2u + 3u - 5w = 2(2, -7, 1) + 3(-3, 0, 4) - 5(0, 5, -8)
= (4, -14, 2) + (-9, 0, 12) + (0, -25, 40)
=(4-9+0, -14 +0- 25, 2+12 +40) =(-5, -39, 54)
4.4 Let s = (2, 3, -4) and v = (1, -5, 8). Find: (a) utu, @) Su, and (c) -u.
I(a) u+u= (2+ 1, 3-5, -4 +8) =(3,-2,4)
(6) Su =(5"2, 5"3, 5"(-4)) = (10, 15, -20)
(c -y=-1+ (1, -5, 8)=(-1, 5, -8)
4.5 Let u and v be the vectors in Problem 4.4. Find: (a) 2u -3u and (b) 3u +Su.
I (a) 2u -3v =(4, 6, -8) +(-3, 15, -24) =(1, 21, -32)
(6) 3u+5v= (6,9, -12) + (5, -25, 40) = (11, -16, 28)
4.6 Let u = (3, -2, 1, 4) and v= (7, 1, -3, 6). Find: (a) u +v, (6) 4u, and (c) 2u - 3u.
I (a) u +v=(3+7, -2+ 1, 1-3, 4+6) =(10, -1, -2, 10)
(6) 4u =(4.3, 4-(-2), 4" 1, 4"4)=(12, -8, 4, 16)
(c) 2u - 3u = (6, -4, 2, 8)+ (-21, -3, 9, -18)= (-15, -7, 11, -10)
4.7 Define the zero vector 0 in R".
Ine Zero vector is the vector whose components are all zero., i.e..0 =(0,0,. ) 115
 CHAPTER4
vectors and row vectors, with examples.
betweencofumn
components are writen
116
Erplainthe
ditterence

vetor u is
2
vector whose vertically, ez,
AColwmm
4
and

and four components,

respectively. A row
three.two. givenin the previous
vector is a vector whose
problems. (Vectors will
are
column
vectorswTth horizontally, as
written
normaly be witten s
are
components
roR vectors.) -6

Coapute
components:
corresponding
I(a) Add

vectors have different numbers of components.

defined Sincethe column
sumis not by the scalar:
(b) The
Multiply each component -3"4 -12\
lc)
-3.-5 15
-3-6/ 18/

the sign of each component:

component by -1; i.e., change
(d) Multíply each

4.10 Compute:

then the vector addition:

IFirst perform the scalar multiplication and

(a)

4.11
(6)
0-949(9-(00-0
Find zand yif (x, 3) =(2, x+y).
17
19

Since the two vectors are equal, the correspondingcomponents are equal to each otner:

X=2, 3=x+y
Substitutex =2 into the second equation to obtain y=1.
4.12
Thus x=2 and y= .
Find x and y if (4, y) = x(2, 3).
Multiply by the scalar xto obtain (4, y) =x(2, 3) =(2r, 3x). Set the corresponding components cqualto cach
other: 4 = 2r,y =3x.
Solve the linear equations for xand
4.13
Find x, y, and z if (2x, 3. y) = y:x=2and y o.
(4, x + z, 2z).
 VECTORS AND MATRICES 0 117
I Set corresponding components cqual to cach other to obtain the system of cquations
2x = 4, 3= xt z, y=2z
and y =2.
Then x= 2, z = 1,
-1) = (1, 4).
4.14 Find x and y if x(1, 1) + y(2,
and y and then add:
First multiply by the scalars x
x(1, 1) +y(2, -1) =(, x)+ (2y, -9)= (x +2y, x - y) = (1.4)
to each other to obtain
Now set corresponding components cqual
x+ 2y = 1, x-y=4
-1.
Solve the system of equations to find x =3 and y

4.15
Evalain what it means for one vector to be a linear combinationof other vectors
Avector v is a linear combination of vectors u,, uz, ...,um if there exist
U= 4,u, t a,u,+ + amum: Scalars a1, a2, : a,, such that

4.16 Wite v =(1,4) as a linear combination of u,= (1, 1) and u, =(2, -1).
We want to express v in the form
V= Xu,t yu
with x and y as yet unknown. Thus we require
(1, 4) = x(1, 1)+ y(2, -1)
From Problem 4.14, x=3 and y= -1. Thus,substituting into (1),we have v=3u. -u
4.17 Write the vector v= (1, -2, 5) as a linear combination of the vectors u, = (1, 1, 1), u,= (1.2 3) and
u, = (2, -1, 1).
We want to express v in the form v = Xu, t yuz t Zu, With x, y, and z as yet unknown. Thus we have

O-0-049-:
x+ y+2z\
x+2y - Z
\x+3y + z/
(It is more convenient to write the vectors as columns when forming linear combinations.) Setting
corresponding components equal to each other we obtain
x+ y t 2z = 1 xty+2z = 1 x+y+2z = 1
x+2y - z=-2 y-3z =-3 y-3z= -3
x+ 3y + z= 5 2y - z = 4 5z = 10

The unique solution of the triangular system is x = -6, y =3, z =2; thus v = -6u, + 3u, +2u,.
4.18 Write v =(2, 3, -5) as a linear combination of u, = (1, 2, -3), u, = (2, -1, -4), and u, =(1, 7, -5).
Find the equivalent system of equations and then solve. Thus
x+2y +z
2r -y+7z
|-3x -4y -5z)

Set corresponding components equal to each other to obtain

x+ 2y + z= 2 x+2y + z= 2
x+2y + z= 2
-Sy + 5z=-1 -Sy + 5z =-|
2x - y +7z = 3 0= 3
-3x 4y - 5z = -5 2y - 2z = 1 as a lincar
be written
The third equation 0=3 indicates that the ucannot
system has no solution. Thus
combination of the vectors u,, uz, and u3.
 VECTORS AND MATRICES 119
... ,0) = 0.
= (0u,,Ou,, .. Ou,) =(0, 0,
. . . u,)
. .k"0) =(0, 0, .. .,0) =0.
(a) Ou = 0(u, , u;,
..0) = (k 0, k " 0,
(0,0,
(6) k0= k
Show that (-1)u=-u.
4.28 previously proved properties,
IFrom the w+-) =0= 0u =(1+(-1))u = lu +(-1)u=u+(-1),
sides.
follows upon adding u to both
and the result
Products, Norms
Dot (Inner) operations:
subsection considers the following vector
This
Dot (Inner) Product: u "V=ab,t azb,t +a,b, = ah
i=|

Norm (Length): |lu|| =Vu"u=Va; t a; t + a?

and v = (b,, b2, . .
b,) are vectors in R".
where u= (a,, a,) . .. , a,)

4.29 Campute u- vwhere (a) u=(2, -3,6), v=(8, 2, -3);(6) u=(1, -8, 0, 5), v= (3,6, 4).
I(a) Multiply corresponding components and add:
u"v=(2)(8) + (-3)(2) + (6)(-3) = 16-6- 18 =-8
(b) The dot product is not defined between vectors with different numbers of components
4.30 Compute u"vwhere: (a) u=(3, -5,2, 1), v= (4,1, -2, 5);(6) u= (1, -2,3, -4), v=(6.7.1. -2).
I Multiply corresponding components and add:
(a) u"v= (3)(4) + (-5)(1)+ (2)(-2) + (1)(5)= 12-5-4+5 =8
(6) u"v =6-14 +3+8=3
4.31 Let u =(2, -7, 1), v=(-3, 0, 4), and w=(0, 5, -8). Find: (a) u"u, (b) u "w, and (c) v" w.

I Multiply corresponding components and then add:

(a) uy=2"(-3) +(-7)"0+1 "4= -6+0+4= -2
(b) u" w=0-35-8= -43
(c) u W=0+0-32 =-32
4.32 Suppose u=(3,2, 1), v = (5, -3, 4), w= (1, 6, -7). Find: (a) (u + v)"w, (b) u " w+v"w.
I(a) First calculate u + v by adding
corresponding components;
utu= (3+5,2-3, 1+4) =(8, -1, 5)
Then compute (u + v)" w = (8)(1) + (-1)(6) + (S)(-7) =8-6-35 = -33.
b) First find u w=3+ 12-7=8 and
U"w5- 18 - 28 =-41. Then
u" w + U"w=8 41 = -33
Note: As expected from Theorem 4.2(i) both values are
equal.
4.33
Let u =(l, 2, 3, -4), v= (5, -6, 7, 8): and k =3. Find: (a) k(u"v), (b) (ku)
U, (c) u"(KU):
I (a) First find u "y=5- 12 + 21
-32 =
(6) First find ku = (3(1), 3(2), 3(3), -18, Then k(u "v)=3(-18) = -.
3(-4)) = (3, 6, 9, -12). Then
-54
(ku)"v= (3)(5) +(6)(-6) +((9)(7) + (-12)(8) = 15- 36 +63 - 96 =
(c) First find ku
=(15, -18, 21, 24). Then
96= -54
u (ku) =(1)(15) +(2)(-18) +(3)(21) +(-4)(24) =115 36 +63 -
Note: As expected from Theorem 4.2(ii) and (iii), allthree values are equal.
Find: (a) |l|| if u =(3, -12, -4), and (b) Ivl| if v =(2, -3, 8, -5).
 by
v w by ftheir each
are
procs of of
component v
multiply is,
component that to perpendicu
The
V109
V109'
V109 zero,as
u. first -3 defined
as each each (c)
|wl|. IS
direction Hence, ifdot
product
divide Divide
=102. 8 is
-3. and R'
16=169 same f.. form.
25 =3, Then 93. |lv||, in vectors
6 v
+\+
4+9+64 k the multinle = normalized and
adding: =v":
=9+144
+ obtain in 169. 64 8 (6)
16+4+9+
vector 1 16= -3
-2 12u||12v| ||u,components. u two
abOVe.
perpendicular)
scalar vectors
and to the
||v||= 39 +9+ = (a)
u +(-4} +8
+(-5}' ||w||
=
V55. unit same 12u When
of obtain )= positive statemen
nonzero
:30 -144 8= Find:
+(-12).
components 5). a
1S 8). the =
the=3V6
Wh
to ||vl=2+(-3)²- -2, +k'
hence k, solve
1
a
IS
= +
-3, 12+(-3)+(-4)'
(-3)'
4
have
û
-8). of
u.
and
(ortwo
the 3).
add V= -2, V54
=(1, ü will and 5, squares orthogonal
u
with -2,.
of between
the |(a) then
3lull'=
squaring 55: Now Since
(4, + W

= u (0, 1= product
(1,
+(-2)' of Then = agrees
u
where30. =
and W
= v, multiple 109 w the+

=9+1+4+Jull'=1'+k'+normalizing
(-2)+5' =
-5). 25+16
û w
as and of 49 be dotangle w
*uby v
13.
component
of
+
V39 k'=
that v.) since
Show
(6)
direction obtain
-4);||v|=y"y=
w"w=42 -3). +9=
4),
sum +
+1=V4=5
R.
to
saidthethe
This and
1),
4, vector obtain -1/4).positive
8, 64 Vg9 physICsV=0.
||u|=
u =V169 V102. -2,
1,(-3,
=lu|| vector. same
-3, to
13 = to
(6, + =(-3,0,
= ||12v||2=36 of +0+16=
the V25
= in are is
64 vectors
-4,

4CHAPTER120find lull=each = |w|'= that nonzero

calleda
unit
the =
(12, =
V169
llwl'V93 2/3, any 12v
(1/2, andfractions: v
root
+(-7) +
25
R" where (3,
in in so =
u

|v| such 1), square + orthogonal v that8 and v

ThenSquare Then w=
v find find=
First Determinek
f have
i|w||
û
is vector
is in
isû (a) First lul|
= Firstl|w| v= v
that
-7, V2 V9 V0
the = = =
and cos
u
(Note=90°,
lul
1),
4,
an ûfinding Moreover.,Normalize: by Normalize"clear" (2, lul| w|| Vectorsu"y=\ul (5,
Note = Take lv|| u"y=0. =
(6) Find We I
be
v
The (a)
I
(b) u
Let (a) (b) (c)
Define other, u
Let
| Let 'of I to I

4.35 4.36 4.37 4.38 4.39 4.40 4.41 4.42

 C ltlhatce)
B. C
Libtary VECTORS AND MATRICES 0 123

Prove Theorem 4.4

andthe other properties of the inner product,
4.54 Cauchy-Schwarzincquality
I Bythe + vl|'(u + v)"(u t u)= u"u -+ 2(u y) + v"u

sides yields the desired inequality.

square rootss of both
Takingthe
lawe
norm in R" satisfies thc following
Prove that the ||u|| 0 iffu=0
4.55 u, ||u|| 0; and
INJ: For any vector any scalar k, lJkul| Jkl llul!
0N.J:For any vector uand v, llu + vl| s ||u|| + |vll.
INJ: For any vectors u and
Theorem 4.2(iv) and the fact that |u|| = Vu w: (N,] is
I[N]follows from
need onlyprove that [N,]holds.
Suppose u= (u,, u,, . u,) and so ku
=(ku,, Minkowski'
kuz, ...sinequality.
,ku,). ThenHence we
|kull'= (ku)' +(ku,)' +..., +(ku,)' =k' u; +
(N,.
Taking square roots gives
42 MATRICES, MATRIX ADDITION, AND SCALAR MULTIPLICATION

This chapter uses capital

letters A, B, C,... to denote matrices andlowercase letters a, b, c, x,y, . . . to denote
eoslars IInless otherwise stated or implied, scalars will be real numbers; in other words, the matioenoe

4.56 Let Abe an arbitrary matrix (rectangular array of scalars), say

a12
A=
a21

\am am2 amn

lal What are the rows and columns of A? (b)
What is the size of A? (c) When are two such matrices
egual2
I(a) The mhorizontal n-tuples
(ai|, a12, ... , ain), (az1, a, - , azn),. ..,(ami, am2
are called the rows of A and the n vertical m-tuples
amn)

a12
l22

\aml \amn
its columns.
(6) Since A has m rows and n columns, it is said to be an m
by n matrix, written m Xn. The pair of
numbers m and n is called the size of the matrix.
(c) To matrices A and B are equal, written A = B, if theyhave the same size and if their corresponding
elements are equal.
4.57 The notation A = (a) is used to denote the matrix in
Problem 4.56 (or A = (a)mxn if the size is not
understood). What is the significance of the first subscript, i, and the second
subscript, ?
ITne scalar a, is the element of A in row iand columni Thus the frst subscript tells the row of the eiem
and the second subscript tells the
column.
4.58
Find the rows, columns,and size of the 2
matrix A= 5 6/
The rows of Aare the horizontal lines of numbers; there are two of them: (1, 2. 3) and(4, 5,6). The
columns of Aare the
vertical lines of numbers; there are three of them:

The size of Ais 2 x 3

(read: 2by 3), the number of rows by the number of columns.
 4
are cqual?
CHAPTER
matrices,ifany,
D following
124

9
the
Whichof c-(i D-G).
B=F
only
Thus
entries.
coesponding

the system offequations

entriesto obtain
corresponding x-y=1, 2z + w =5, Z- W =4
/Equate ity=3,
equationsissr=2, y =1, z =3, w=-1.
systemof whose
ofthe simply 0, is the matrix
Thesolution
denoted by Omx Or
matrix,
are elements all zerO.
Display the
Xnzero
matrices.
The m and1 x4zero
4.61
2x2,2x3, scalar0:
eachentrybethe
let
/Simply (0,0,0, 0)

=(a,)is the m
Xn matrix -A=(-a,). Find the negatives of
matrix A
an mXn
4.62 The negative of 1 -3 4
-5 0 -8/

eachelement:
Take the negative of 3 -4 -7)
-(-3) -4
-(-5) -0 -(-8)/ 0 -a-()
frequently denoted bu
matrx or a row vector and is
matrix Awith only one row is called a row with onlv one
it must be one. Analogously, a matrix B
4.63 A
A=la,a, . .. ,a,); we omit its first subscript since
vector and is frequentlydenoted by
column is called a column matri or a column

b,
B=

(nere we omit the second subscript since, again, it must be one). Discuss the difference, if any between the
following

u= (1,2, 3) and

IViewed as
vectors in u
they have different sizes.R', and v may be considered equal. However, as matrices, they cannot be equal,for
Matrix Addition and Scalar
This
B= (b)subsect
are
ion considers Multiplication
the following
matrices with operations of
the same size and k matrix addition and scalar multiplication where A=(4,) and
is a
a tbu aytb2 scalar:
A+B= ain t bin ka,
az2t b kat kaiz
kaz.
kaz1 kaz
the \amitm Qm2t bm2 and kA =
That is, A+ Bis
amn t b.m
multiplying each entry of Abyobtk.ained by adding
matrix \kami ka,m2
corresponding entries from Aand Band kA is the matrix obtained by
 VECTORS AND MATRICES 0 125
3 0 2
and 2
where: -4
Find A+ B
1+3 -2+0
entries: A + B =
3+2\ 4 -2
corresponding 5+1 -6+8)
Ic) Add since the matrices have different sizes.
is not defined,
(6) The sum

4.65 Find C+Dwhere:

-1 2 -3

5
and D=( 3 1 -1) and D=((3 -S
0 -2
IAdd coresponding elements:
1+1 2+(-1) 3+2
(a) C+D=0 5+3 6+(-5)/
(1+3 2+(-5) (-3) +6 4+(1) -3 3

(0) C+D=lo+2 (-5)+0 1+(-2) (-1)+(-3)/= -5 -1

-2
4.66 FiFd 3A and -54, where A= 5

Multiply eachentry by the given scalar:

34=(3"1 3"(-2) 3-3
34 3"5 3-(-9)-(123 -615 9
-18)
-5. -5"(-2) -5"3
-SA=5.4 -5-5
10 -15
30

4.67
Compute: (0) ).0-2 -3

-2)"1 (-2)·7
(-2)"2
\(-2)-0 (-2)"(-1) / 0 2/
4.68 The difference, A -B, of two matrices Aand B of the same size is defined by A-
B=A+(-B), Find A-B if
" and (2 -3
B= -2
-5 -?
A- B=A t(-B)=
2 3 2

4.69 Find 2A -3B, where A -2 3 3 0 2)

5 -6) and B-(
I First perform the scalar
multiplications, and then a matrix addition:
-4
2A- 3B= 28 -4
10
7 -36/
21 -3
(Note that we subtracting. This usually
avoids erTOrs.) multiply Bby -3 and then add, rather than B by 3 and
multiplying
4.70 IfA=(?3 s -2 1 -2 find 3A + 4B
- 2C.
0
and C=(
 CHAPTER4
128 D matrices Aand Bin Problem 4.77.
BA ofthe
product
Findthe
4.78
-2
defined.
1) andB =/!
(2
ABwhereA =
4

Findtheproduct product AB is defined as a 1 x

4.79
Bis 2 x 3, the 3
Components. Tox2
I Since Ais1
and the components of AB, multiply the row of Aby each column
obtain or
roW
of B: matrix,
vector with 3
B=2 L)-(2+4,,-4+5,0- 3)=(6, 1, -3)
2 -1
-2
1 0and B:
A=
theproductABwhere 4
4.
Find -3
product AB is defined and is a
X 2 and Bis 2 X3, so the 3x3 matrix. To
theINow
product 3
Ais matrix AB, multiplythe first row of Aby each column of B, respectively:
obtain the first ro%
/2-3 -4-4 -10 + 0\
-1 -8 -10\
4
|-3
To obtain the second row of the product matrix AB, multiply the second row of A by each column of B.

respectively:
-1 -8 -10
-8 -10\
*0 -s+0
1+0 -2 +0 1 -2 -5

To obtain the third row of the product matrix AB, multiply the third row of A by each column of R
respectively:
-1 -8 -10 -1 -8 -10\
1 -2 -5 1 -2
-3 \-3+12 6+16 15+0/ 9 22 15
Thus
-1 -8 -10
AB= 1 -2 -5
9 22 15)
4.81 Find the product BA where A and B are
the
matrices of Problem 4.80.
I Since b is 2X3 and Ais 3 x 2, the
multiply the first row of Bby each product BA is defned as a 2 X2 matrix. To obtain the first
column of A, respectively: roWOT De
/1 -2
3

To obtain the
second row of BA, multiply the second row of B by each column of A, respectively:
-21
Thus -3+0+0)-0)
Find AB
where A=( -2
3
and B=
2
1
-1 0 6
-5
 +
ACis
matrix A
/-entryofthe
ase,(anba, tauca) (1)
lHencethe 2
matrixAD-A(B
+ C) is
/-entryof the
otherhand,
the
S and, a,(b,a t e,) (2)
Onthe
ofthe distributive lawin the scalar field; this proves (ii).
equalbyvirtue
are
sides of ( )and(2)
Theright BA + CA.
4.6(ii):(3 +CA between left and right multiplication in the fheld of
ProveThcorem 4.90.(Thereis no distinction
4.91 Problem
The proof is as in
|
sCalars.)
k(AB) - (kA)B- A(kB),
4.6(iv):
Prove Theorem
4.92

NSPOSE OF A MATRIX
4.4 TRA on matrices
where the transpose of a matrix A, written A, is the matrix
considers the transpose operationcolumns:
Thussection as
rows of 4, in order,
obtained by writing the a, b, C

b, a, b, Cz
b, b,

\C
C. b, C,

where a = a.. for all iand i

matrix, then A=(a) is the n x m matrix
In otherwords. if A = (a,)is an m X n
3 find 4 and (4')".
4.93 Given A=(6 - 7 -8
the rows of A' as columns to obtain (A'):
Rewrite the rows of A as columns to obtain A', and then rewrite
3
A'=3 -7 \6 -7
5 -8
(As expected (see Theorem 4.7(i)], (A") = A.)
1 2 3
4.94
Transpose: la)( and (b) -4 -4 -4
5 6

I Rewrite the rows as columns in (a) and the columns as rows in (b):
-4 5\
(a) (6) 2 -4 6
3 -4 7,

4.95
Find u', v', w' for the row yectors u = (2. 4). =(1.3. 5), and w =(6, 6, 0).
I Thetranspose of a row
vector will be a column vector:

w=6
 132 D CHAPTER4

The
transposesof
the column
vectors u=
.-)-)--)
transposeof a columnvector will be arow vector: u =(1, 1), v' =(2,4, 6), w' =(-5,-6.1
the
Find

defined|for any matrix A.

AAa and A' are
matrices
IfAisan mXn matrix, then ATis an n X m matrix. Hence AAT;ie Aefned as an mX mmatrix, and A'A.
the
Showthat
4.97 matrix.
n Xn
definedasan

A'A,whereA=( -1
FindA4and Aas
columns:
4.98
rewritingtherows of
IObtainA'by 1 3

whence AA-(1 -

1+9 2-3 0+12\ 10 -1 12

2-3 4+1 0-4 -1 5 -4
\0+ 12 0-4 0+ 16, 12 -4 16)

Find (AB)".
4.99 14) 39
AB= / 20+14)
15 +24 0- 28/
-( -28)
and so
(a)'-( 14 -28

Problem 4.99.
the matrices in
4.100 Find AB for
We have 5+0 -6+21)
and B=() and so A'B'=10+0 -12- 28/ \10 -40/
Note that (AB) *ATB".
4.101 Find B'A for the matrices in Problem 4.99.
15 + 24 39\
/5- 12
\0+ 14 0- 28/ )-67 14 -28/

/ (As expected (see Theorem 4.7(iv)], (AB) =B'A".)

Theorem 4.7: The transpose operation on matrices satisfies
() (4+ B)' =A" +B" (üi) (kA)" = kA" (k a scalar)
(i) (A) =A (iv) (AB)" = BA
4.102 Prove Theorem 4.7(i): (A + B =A + B".
I IfA=(4,) and B=(b,), then a, +b, is the ij-entry of A+ B; hence 4y +b, is the ji-entry (reverse order) of
(A+ B)'. On the other
A' + B'. Thus (A + hand, a,, is the ji-entry of A' and b, is the ji-entry of B'; so ay + b, is the ji-entry of
B)' = A" +
4.103 B',since corresponding entries are equal.
Prove Theorem 4.7(i):
ISuppose A= (A")' =A.
c, =b, =ay, (a,), A" =(b), and (A") hence
Thus =(c,). Then, for every iand j, we have b, =aji and interchange of
rows and columns
4.104 Prove
(A')=A,
is since
equivalent to no corresponding elements are equal. (In other words, a
double
Theorem
IIf A= 4.7(ii): (kA) kA where k is
interchange.
(a,), then ka, is the a
ij-entry of kA, and soscalar.
ka,, is the ji-entry (reverse order) of (kA). On the other
 4.142 4.141 (d) (c) (b) (a) This 4.6 4.140
4.139 4.138 4.137 4.1.M0
is A(i>).
s A D=(d, A ) SPECIAL
section
CHAPTER
140 4
Find (T te symmetric =(a,)
Wriorthogonal
IS Put I IBy Show RA= IfA Show kAof
.
I IfA and invertibleCinsot
AB
common
the out is
considers
is that that Sin ce is Ib) (a) I (4B) Form Apyply
where diagona, TYPES Problem AB=. is A
upper invertible,
k Thus for The Thus We the the
given diag the invertible, is
practice if A'
A' (lower)
if
0, th e n. have 'B block Gnus ian
A scalars (3, the
4.operations
139, invertible k=show proof B
We
diag

-3,5)
(2,
=

omit
to
the
on
diag =A, =A, denoted
-7,2),
i.e.,
i.e.,triangular
if
a,
following OF
SQUARE
by
square the
B
is of A'
inversion
Hence then
(A,A,A,A,1)t'h=eor(eAm ,A, induction
there and if exists. kAholds
1/k that prove isbyis
it the
is
'A '4
(6) -6
0
matrix
nlMgorithm
blocks
-70
0 3
diagonal, (4,
-5),
if
AA' a,= itsif =
D
inverse invertible,exists only
is
Then
invertible
is
for
true inverse
for on (B
A,A,A, 0 00 (( to
and 0 diag I Ce
B
patterns
or 20
0 and A'A al
foelements
r matrices:
MATRICES
(d,,, of
and a if
A' (kA)(k
n
+ n+1. n.
AB,
(4B)X(B
of
'A is and
)
= with matrix is 1. For
diag Os
diag =1. i
and d,...
below
transposition
that
A'; with
invertible. Accordingly, Using n=is, (AB)
when that invertible
'A 0 0|3 row
(7,4, of elsewhere: (6, j. inverse such B -'A-)= 1, )=
4 2
reduce 6
k 4)
Os -3,0, (A
(above) ,
is #0,A,)A,.J= part (AB) we B(A == B (A,A,and 6
6). as is d,.), B=(A'). B". that (kk (a), have A(BB)A' , cchelon
form: to
in skew-symmetric commute; withtheoremthen. B-'A-!.'=
the 1). if The AB 'A)B
the its )(AA)=1"l=1. inverse A,=A.Suppose wen 0
third nondiagonal
diagonal converse BA = A,(A,A,. have = -
0\
/6 But that B A,)A,'.
B'IAlA=
matrx.) = k holds \00
-3 or B=A; is, follows(AB)' Then
I.
=
'A-!.
antisymmetric are (A")= ·A,)=A,.A, for
-B= AA
2
0 elementsall every
zero, hence from inverseHence >1 =| A. A,'
(A-'). positive and I
i.e., are (A-') the k'Ais k
(BA)" = the
if al fact theorem
A* ifzero. that
a, = the
=-A.) = (A A,'A; holds
0 (aT and
when so
i<j
 D CHAPTER4
matrices of. orders 2.,3, and 4.
142 lowertriangular
generic diagonal:
Exhibitthe abovethe
put0s
4.148
cachcase
n |bin
b1 b2 C Cn
Cy C2 C
\Ca CA). Cay
Matrices Cu
Skew-Symmetric
and ifcachofthefollowing matrices is symmetric or skew-symmetric:
Simmetric
Determine

5 -2 4 -7 1
4.149
3 (6) B= -7 3
(a) A = -5
2-3 0 2

elementsare all 0and

the diagonaland -3, are negatives ofsymmetric elements
hence (mir or IfMmages in the diazr
inspection. 3 each other:
|a) By -2and 2, and A
Sand-5. the symmetric elements,
-7 and -7, 1and 1,
and 2and is
(b) By
inspection,
symmetri.
Cis notsquare, Cis neither
symmetric nor skew--symmetric.
2, are
skeequalw-sym;mehencetric. Bis
(c) Since
matrices is symmetric or skew-symmetric:
of the following
Determine if each
4.1S0

(a)
symmetric.
D; hence D is
D'#=tE.
(b) ByWeinspection
I(a) see that E' Accordingly, Eis neither symmetric nor skew-symmetric.
hence Fis skew-symmetric.
(c) By inspection, F' = -F;when 0 is square.
(d) Both, since 0' =0= -0
/S 2

4.151 Find . y, z, tif A=y z -3 is symmetric.

4 -

| Equate symmetricelements (mirror images in the diagonal) to obtain x =4, y =2, t= -3. The unkn
on the diagonal, is indeterminate.

4 x+2
4.152 Find x and Bif B =
\2r-3 x+1, is symmetric.
ISet the symmetric elements x+2and 2r+3equal to each 4

4.155 Suppose A= (a,) and B= (b,) are symmetric. Show

other to obtain x=5; hence
B-\
that (a) A + B is symmetric, and (b) kA 1S SYmne
(a) If A+B=(c,), then c,, =
(b) If kA=(c,), then c, = a, + b, = a, + b, = C,i
ka, = ka, = Cij
4.154
Show that AB need not be
symmetric, even though Aand Bare symmetric.
I Let A=2 2/and
2) R
14 17
5 6/ Then AB =
Theorem 4.8: Let Abe a 23 28 is not symmetric.
square matrix. Then: (i) A+
where B=(A+
A')is A' 1s symmetric; (ü) A-A' is skew-symmetric; (ii) A
4.155
Prove Theorem 4.8(i). symmetric
and C = A-A) is
skew-symmetric.
I (A+
A'y =A'
+(A'Y =A' +A =A+A'.
 14%

M) A A A')
Theorem 4
Prove
4.IS6

Theorem 4
ni)
A-# Candthat ny muiple
4. I67
Prove
toltow thefct that
fromsymmetric)
I The proof (ekew
symmetric
mattix is symmetre utist
um of a symmetrie matrix Mand a skew
Write A- asthe
4.158

ICalulate

Then
A" is symmetic
symmnetric. Show that A' and, in general,
Suppose Ais
4.150
A'. Alo, by induction, (A")'-AA"
(A)'-(AA)' A'A'-AA
matrix and P any n m matrix, Show that P'APis alo ymmetric.
n-square symmetric
4.160 Let Abe an
- P'AP.
I P'AP)' - P'A'(P'y'
OrthogonalMatrices
/1/9 8/9 -4/9

4.161 Show that A = 4/9 -4/9 -7/9 is orthogonal.

8/9 1/9 4/9

We nced only show AA":

8/9 -4/9\ 1/9 4/9 8/9 1+64 + 16 4- 32 + 28 8+8-16
1/9 1
8/9 -4/9 1/9 4-32 +28 16 +t 16 + 49 32 -4- 28
AA' 4/9 -4/9 -7/9 81
8+8-16 32 -4 - 28 64 + 1+ 16/
8/9 1/9 4/9, -4/9 -7/9 4/9/
81 1 0 0\
0 81
81
0 0 81 \0 0

of vectors in R".
4.162 Define an orthogonal set and an orthonormal set
orthogonal set of vectors if the vectors are pairwise orthogonal, that s,S,
an that
I The vectors u,, uz, ..., u, form orthonormal set if they are orthogonal and if they are also unit vectors,
4, 4, =O for ij. They form an Kronecker delta., the condition for orthonormality is 4, 4
Oy
if u, u, = 1 for each i. In terms of the

4.163 Show that

b, ba b orthonormal
C2 Cs form an
and u, (C,,Cz, C,)
is orthogonal if and only if its rows u, =(a,, 4z, 4). uz =(b, b,, bi),
set.

I IfAis orthogonal, then I 0 0

AA = b, b, b,
C2 C,/
b,
b, C/
\o ) -
 14 definitionof
The
general
det (A)=)*sgn (o)
TOlows:
Remark2: 4y4
over all
permutations o
number of. (i,:aren} of {1, 2,.
wherethe
sum is
taken
even oranodd interchanges rcquired to chanoe
,n). Here
one
accordingas an
bricllytreatsuch determinantslaterinthesection.) Permutations are introduced in Chapter 8. sg(o) eawalh
whose proofs lie
theorems, beyond the scope of
The
followingtwo this tex,
Remark3: play an
theory.

Theorem4.10:
Forany
two
n-square matrices AIand B, we have

det (AB) =det (A)" det (B)

Impotant tole in

determinant function is multiplicative.)

(That is, the
invertible if and only if it has a nonzero determinant.
A matrix is
Theoremm 4.11:
One and Two
Determinants of Orders
(-6), and det (t+2).
4.172 Find det(24), det
the scalar itself: hence, det (24) =24, det (-6) =: -6, and det (t +2) = +2.
The determinant is
Show that the equation ar =bhas a unique solution if and only if det (a)+0.
4.173
solution if and only if a #0. However, det (a):
IThe equation ax = b has a unique
det (a))#0. (Analogously, the matrix equation AX =)=a. Thus ax =b has a
unique solution if and only if Bhas a
unique solutio
and only if det (A) +0.]
4.174 Give amnemonic for the evaluation of the determinant of order two.

The determinant is equal to the product of the elements along the plus-labeled arrow minus the product n
clements along the minus-labeled arrow. [There is an analogous scheme (Problem 4.185) for determinant
order three, but not for higher-order determinants.]
4.175 Compute the determinant of each matrix:

I(a)=(6)(3)- (4)(2) =15 -8=7 2

12

(c) 13
(b) Li-26)
-4 - (1)N-4) =12+4= 16
15+8=23
4.176 Compute the
determinant of
4

-2 -8-5= -13, (6) 15 -0= 15, (c)

la b = ad - bc

4.177
Find the determinant öf: (a) A
=(O a Q+
a

I(a) det
(A) =
a+ b =(a -
b)(a + b) -((a)(a) = -b²
(b) det (B) |-5
=\1 +3=(- 5)(+ 3) +7=2-15
+7=-2 -8