UNIT 3 FUNCTIONS OF SEVERAL .

VARIABLES
Structure
3.1 Introduction
Objatives
3.2 The space Rn
Capxian Products
Algebraic Shuctrne of Rn
Distance in '
R
3.3 Functions From Rnto R m
3.4 Summary
3.5 Solutions and Answers

3.1 INTRODUCTION

In the fust course of calculus and the first two d t s of this block we have studied the
concepts of limit, continuity and differentiability of real-valued functions of a real variable,
i.e., functions whose domain and range are subsets of R,the set of real numbers. In Block 2
we shall study these concepts for functions of several variables, i.e., those functions whose
dom& is a subset of Rn,the Carhisian product of n-copies of R.These functions arise
naturally in various contexts. For instance, the insurance premium is a function of a large
number of parameters like the sum insured, the age of the insured person and life expectancy.
Similarly the price of a commodity is dependent on a number of factors like cost of
production, permissible profit margin and state taxes.
You have seen that a knowledge of the algebraic structure of R and familiarity with the
I
properties of the distance ( x - y between two points x. y of R is necessary to study the
notions of limit and continuity of functions of one variable. The same is true for functions
of,several variables. Therefore, in this unit we first defm R4 and describe its algebraic
saucture. We then introduce the notion of a distance between two points of Rnand deduce its
elementary properties. We end this unit by defining a function of several variables, and by
giving various examples of such functions.

Objectives
After reading this unit, you should be able to :
I. defm a real Euclidean space of dimension n

I give examples of real-valued and vector-valued functions of several variables.

3;2 THE SPACE Rn ..

p (G,
yo) We have mentioned in the introduction that in this unit we are going to study functions
whose domain is a subset of Rn.But what is Rn? In this section we shall define Rn,and
study its algebraic structure. We'll also study the distance function in Rn.But let us start
with the definition of Rn.For this we'll need to define Cartesian products of sets.
I
3.2.1 Cartesian Products
Flg. 1
Let X and Y be two noncmpty sets. By (x, y) where x E X and y E Y, we denote the
The word Cartdam product ordered pair whose first member or coordinate is x and whose second member or
~amfffromthellame~fthe
French mamematiciw Renc' coordinate is y. Two ordered pairs (x,, y,), (x,, y,) are said to be equal, i.e.
Descartes (1596-1650). who first (x,, y,) = (x,, y,) if and only if xl = x,, y, = y2. YOUare already familiar with this concept
thought of representing points in While studying coordinate geometry you must have represented a point P in the Cartesian
a plane by ordaed pairs of plane by (x,yo), where xo is the abscissa and yo is the ordinate of P (see Fig. 1). Clearly the
numbcrs.b point (x,yd &different from the point (yo. x,,) if x,,# yo. Thus you know that a point in
,the plane is represented by -an ordered pair (x.- y), where x and y are real numbers.
I
 Functions of Several
Note that the ordered pair (x, y) is different from the set (x, y 1, because the ordered pair (x, y) Variables
is different from the ordered pair (y, x) if x # y, while the sets (x, y ) and ( y, x 1 are equal.
The set of all ordered pairs (x,y) where x E X, y E Y is called the Cartesian product of
the sets X and Y. We denote it by X x Y.
Thus,X x Y = ( ( x v y )I X E X,YE Y).
For example, if X = (0, 1.2) and Y = (0, 1), then
x x y = ((O,O),(O, l),(l,O),(l, 1),(2,0),(2, 1))
IfX=R,Y=R,then
XxYFRxR=RZ=((x,y)I x E R , ~R
E)
is nothing but the Cartesian plane.
We now extend this idea to get a product of n sets.

Let XI, X2........ Xn be any n non-empty sets. By (x,, x2, ...... x,) where xi E X;, 1 I i I n,
we shall denote an n-tuple. Two n-tuples (xl, x2, ......... x,) and (y,,y2.......y,) are equal, i.e.

if and only if xi = yi for all i, 1 I i I n.

The set of all n-tuples (x,, x2.....,xn) where xi E Xi, is called the Cartesian product of n sets
XI, X,, ....... X,. It is denoted by X1 x Xz x ......x X,. Thus,
X1 x X Z x ........x X, = ((x,, x2. x3, ..... x,) I xi E Xi, 1S i L n }
Now,ifX,= ( 1 , 2 ) , X z =( 1 , 2 ) , X 3 = (O},thenwhatwillX, x X Z x X 3 b e ? Youcan
easily check that XI x X2 x X3 = ((1, 1,0), (1, 2, O), (2, 1,0), (2, 2.0)).
If Xi = R for all i, 1 I i I n, then
XI x ...x X, = R x R x .... x R (n times) = RD
= ( (x!, ...... x,) I xi E R. I L i I n)
is called the Cartesian prodbc~ofn-copies of R.
Note that R' = R, RZ.is the Cartesian plane, and R3 is nothing but the set of all points in
3-dimensional space.

Let V denote the set of all vectors 6b in the Cartesian plane where 0 is the origin and P is
any point with coordinates (x, y) in the plane. Then there is a one-to-one correspondence
between V and Rz given by 6b + (x, y). Similarly there is a one-to-one correspondence
between the vectors & (when3 O is the origin and P is any wint (x, y, z) in space) and the
points in space given by 6b (x, y, z). It is because'of these correspondences that
elements of RZand R3 are called vectors. In what follows, we shall call elements of Rn,
vectors and the elements of R will be r e f e d to as scalars. *:
If x = (xI. xz, ,,..., x,) is any point of Rn, then xi is called the i-th coordinate or the
a
i-th component of x.
Now having defined the set RD let us see if we can define any algebraic operations on the
elements of Rn.
. .

3.2.2 Algebraic Structure of Rn

In the last sub-section we have seen that for any integer n 2 1

R n = ( x = ( x,,...., x,)I x i € R, l l i l n ) ,
and two vectors x = (x,, ..... x,), y = (y,, .....y,) .
are equal if and only if xi = yi 1 S i 5 n.
We'll now introduce an algebraic structure on Rn :
The zero-vdctor : The vector, all of whose coordinates are 0, will be denoted by 0 and it
will be clear from the context whether 0 represents an element of Rn or the real number 0.
Advanced Calculus Sum of two vectors: Let x = (x,, ......, x,) and y = (y,, ....., y,) be two elements of *Ei
Then x + y, the sum of the two vectors x and y, is defined to be the vector whose i-th
coordinate is xi + y,, 1 I i 5 n. That is,
x + y = (x, + Y,, x2 + Y2' '..'.., X, + y,)
Using the properties of the operation of addition in the set of real numbers, we can easily
prove the following :
A1 If x, y E Rn,then x + y E Rn.
A2 x + 0 = 0 + x = x for any vector x in Rn.
(Here 0 denotes the zero-vector.)
A3 Given any vector x in Rn,there exists a unique vector y in Rn such that
x+y=y+x=O.
A4 (x + y) + z = x + (y + z) for any three vectors x, y, z in Rn.
A5 x + y = y + x for any two vectors x, y in Rn.
If x = (x,, ......,x,), then the unique vector y mentioned in A3 above is clearly equal to
(-x,, -xz, ......, -xn). We shall denote it by -x and call it the additive inverse or
negative of x.
If x and y are two vectors in Rn,then x - y, the difference of x and y, will denote the vector
x + (-y), where -y is the negative of y.
Scalar Multiplication: Let x = (x,, x,, ......,x,) be any vector, and let a be any element
of R. We define a new vector ax by
ax = (ax,, axz, .......,ax,)
We say that the vector ax has been obtained by multiplying the vector x by the scalar a and
this particular operation is called scalar multiplication in Rn.Clearly 0.x = 0 for every
.
x in Rn.Note that 0 on the left denotes the real number 0 and 0 on the right hand side
denotes the vector 0 in Rn.
The following properties bf scalar multiplication are easy to prove, and we leave them to
you as an exercise. See El).

S2. a (x+y) = ax + ay for every x, y E Rn,a E R.

S3. a(bx) = (ab) x for any x E Rn,a, b e R.
S4. (a+b) x = ax + bx for x E Rn,a, b E R.
S5. ax = 0 for every x E Rn if and only if a = 0.

E 1) Prove S 1, S2, S3, S4 and S5 by using the corresponding properties of real numbers.

You must have noticed that the addition of vectors and scalar multiplication in Rnare
identical with the usual operations of addition and multiplication of a vector by a scalar in
the plane or space (when n = 2 or 3).
pIK - YI*
"
Further, you must be aware that multiplication or division of vectcrs is not defined in the
plane or the space. Similarly, we do not define these operations in Rnfor n > 2.
X

Those of you who have studied the course Linear Algebra, would have recognised that Rnis
a vector space over R w.r.t. the operations of addition of vec:ors and scalar multiplication
defined above.
Now after this discussion of the algebraic structure of Rn,let us define a distance function
in Rn.
p- lx-YI *
Y x 3.2.3 Distance in Rn
Fig. 2
Distance berween x and y is You know that for any two real numbers x, y, the absolute value x 1 - y1 =
I x-yl denotes the distance between the points represented by x and y on the real line. See Fig. 2.
58 Similarly, from your study of coordinate geometry you know thatthe expressicn
 .
4 (x , - x2)2 + (y I - y2)2 represents the distance between two points with coordinates
Functions of Several
Variables
(x,, y,) and (x,, y2) in the Cartesian plane. We define the distance between any two points of
Rn in such a way that on taking n = 1 or n = 2, our distance formula reduces, respectively,
to the two expressions mentioned above.
Definition 1: Let x = (x,, .......,x,) and y = (y,, .....,y,) be two points of Rn,we define
( x - yl, the distance of x from by

I
b
You can see that for n =I. x-y = / 1 is nothing but the absolute value of x - y.
which is the distance between the points x and y .on thereal line.

~1
F o r n = 2 , Ix - = d ( x l - Y , ) ~
+ ( ~ 2 - y ~ ,)which is the distance between the points
with coordinates (x, ,x2)and (y,, y2) in the Cartesian plane.

Those of you who have studied coordinate geometry of 3-dimensions would recognise that
forn=3

Ix-YI = + X I -Y,)~ + (x2 - y2I2+ (x3 - y3I2 is the distance between two pints with
coordinates (x,, $, x 3 and (YI.Y2, ~ 3in) space-
The distance between two points of Rn,defined in this way, has the following properties,
which are easily deducible from the defmition.
k t x = (x,, x,, ...., x,) and y = (y,, y2, ......., y,) be any two points of Rn.Then

You know that the sum of two sides of a mangle in the Cartesian plane or space is greater
than the third. This means that if x, y, z are t e points in Rn,n = 1.2, 3, then

I The same is true for Rnfor n > 3, i.e.

for any three points x, y, z of Rn.But to prove this fact, we have to take the help of
Cauchy's inequality which we shall now state.
Theorem 1(Cauchy's Inequality) : If a,, a2, ....., %, b,, b,, ........., b, are any 2n real
numbers, then

Proof If all the real numbers ai are equal to zero,then there is nothing to prove. Assume,
therefore, that at least one ai # 0. Consider the expression

Clearly ax2 + 2bx + c 2 0 for all real x, and the identity

implies that ac - b2 2 0 as a > 0.

Thus b2 Iac, or
This means

and the proof is complete.

Now for any point x = (x,. ......,x,) in Rn,

is called the nwm or rnedulus of x. Recall that we use the satne terms when x E R,or x
is a vector in the plane or the space. Now let us get back to the mangle inequality.
Theorem 2 (Triangle inequality) : For any three points x, y, z in Rn

lx - YI S I X - zl + l z - YI
Proof : Let us first prove that

Ix + YI 1x1 + Iyl
for any two points x, y in Rn.
Let x = (x,, ....., x,), y = (yl, ....., y,).

in view of Cauchy's inequality.

Consequently,

Ix + YI'~xI'+ 21x1 Iyl + ly12.

Now, if x, y, z are any three points in Rn,then

The set Rntogether with the distance between two points of R. defined above is called the
Euclidean space of dimension n.
We would like to tell you that

(xi - yi)2 is not the only way in which we cap defir.3 distance in Rm.In
fact, there are many other ways in which we can define distance between any two points of
Rnwhich will also satisfy the triangle ine@ality. But Rnendowed with any distance
function different from &heone defined above is not called a Euclidean space. We shall not
deal with any space other than the Euclidean space here.
You know that sets of the type ]a, b[ = ( x E R I a < x c b ) where a and b are real numbers,
oa or - m are called open intervals in R. We now, introduce analogues of open intervals in
the Euclidean space Rn.
Definition 2 :Let xo E Rn and r > 0 be any
. .real
- number. Then the set
 ~ ( w r ) =x( IXE 1
R.. x-xo ( < r ) a
Functions of Several
Variables
is called an open sphere or open ball or open disc with centre xo and radius r.
Remark 1: .i) If n = 1, then S(Q, r) is nothing but the open interval ]xo - r, xo + r[. See
Fig. 3(a).
ii) If n = 2 and xo is the point with coordinates (a, b), then S(xo ,r) is the interior of the
disc in the plane whose centre is (a. b), and whose radius is r. That is,
s or = x Y I - a), + (Y- b12 < r 1.
See Fig. 3 (b).
iii) If n = 3 and xo is the point in space with coordinates (a, b, c), then
S 1
) = ( ( x y z) d ( x - a)2 + (y - b)2 + (z - c ) ~< r ), i.e., the interior of the
sphere whose cehtre is (a, b, c) and radius is r. Also see Fig. 3 (c).

Fig. 3 : (a) S(xg, r) In R, (b) S ( q , r) in R2, (c) S ( q , r) in R3

Remark 2 :In an analogy with neighbourhoods on the real line, we shall call the open
sphere S(xo, r) an r-neighbourhood of the point xo in Rn.By the deleted
r- neighbourhood of x,, we shall mean the set of points

We conclude this sedtion with a few examples and exercises. Go through the examples
carefully and try all the exercises. A thorough knowledge of the structure of Rn will help
you while studying the limit and continuity of functions of several variables in Block 2.
Example 1 :Let el = (1,0, O), e2 = (0, 1,O) and e3 = (0,0, 1). Then we can show that
x = (x,, x2, x,) in R3 can be uniquely written as
x = x,e, + x2 e2 + x3e3.
By definition of scalar multiplication we get
' xlel = (xl, O,O)

If you have done the course on

linear algebra, you would recall
that (e,, e2,ejl fonns a basis of
R3
and therefore xlel + x2e2+ x3e3= (xl, x2,x3) = X. NOWlet us prove that these x,, x,, x, are
unique. Thus; if
Advanced ca~cu~us x = alel + %e, + a,e, ;a,, %, a, real, then
x = (a,, %, q)=,(x,,x,, x,). and therefore
a,= x , , ~ = x , , a , = x , .
The vectors el. e,, e, are called unit vectors along the coordinate axes.
Example 2 :Let x = el + e, - 2e3, y = 2e1- e, + e, where el, e,, e, are the unit vectors
definedinExample1.Letusfind lx + 2 y / , ix + yl.
Now, x + 2 y = e l + e 2 - 2 e 3 + 4 e l - 2 e 2 + 2 e 3

Similarly,

Example 3 :The open disc S with centre (a, b) and radius r in R2 lies in the open square

qnd contains the apen square

S, = { (x, y) ( 1 x - a1 < r / fi I y - bl < r / fi ).,See Fig, 4.

Now let us try to prove this.
If (x, y) E S, then we know that

and therefore,

That is, (x, y) E S,. mis means S c S1.

Fig.4 Now, if (x, y) E S,, then

and therefore,

That is, (x, y) E S. Thus, S, c S.

See if .you can do these exercises now.

E 2) Let e, = (6,,, 6,,, ........,6,). 1 5 i In where tjij is .the Kronecker symbol, (6ij= 0
if i # j, = 1) be n vectors in RD.Prove that any x = (x,,....., x,) in RDcan be
written uniquely as

el, .......,enare called the unit vectors along the coordinate axes.
E3) Lete= (l,O),f = (1, 1)be inR2.
~ i n d l x- yl, )2x - yl, 1x1 wherex=e+f,y=2e+3f.
E 4) Show that the open sphere S with centre at (a, b, c) and radius r in R3is contained
in the open cube

P , = ( ( X . Y . @ ~I x - a l < r , l y - b J < r , ~ z - c l ~ r )
and contains the open cube
 Functions of Several
Variables

Now we shall turn our attention to functions defined on subsets of Rn.

3.3 FUNCTIONS FROM Rn TO Rm

You have already come across the definition of a function earlier (see Definition 4, Unit 1 of
Calculus). Thus, if X and Y are two non-empty sets, then a function from X to Y is a rule
or correspondence which associates to each member of X, a unique member of Y. Here, we
shall introduce a special type of function for which X is a subset of Rn,and Y is a subset of
Rm,both m, n 2 1. If m = 1, such functions are called real-valued functions of n
variables. And it m > 1, these functions are called vector-valued functions of n variables.
More precisely, we have the following definitions:
Definition 3 : i) Let D be a non-empty subset of Rn,the Euclidean space of dimension n,
n 2 1. A function from D to R is called a real-valued function of n variables with
domain D.

ii) Let D be a non-empcy subset of Rn,n 2 1. A function from D to R m (m > I) is called

a vector-valued function of n variables with domain D.
A function of n variables is also called a function of several variables.
Iff : D + Rm,where D c Rn,then we denote the value of the function f a t a point
x = (xI, x2. ....., x,) E D by f(x) or by f(x,, x2, ....., x,).

Now here are a few examples of functions of several variables.

i) For (x, y) E R2,define f(x, y) = sinx + cosy.
Then f(x,y) is a real-valued function of two variables with domain R2.
ii) Let D = [-I, 11 x 1-1, 11. For (x, y) E D define f(x, y) = sin-' x cos-I y. Then f(x, y) is
a real-valued function of two variables with domain D.
iii) For (x, y, z) E R3,set f(x, y, z) = XI + 2) I + z
2
. Then f(x, y, z) is a real-valued
function of 3 variables with domain R3.
iv) Let x = (x,, x,, ......, x,) be any element of Rn.For any j, 1 I j I n, define
xj (x) = xj = j-th coordinate of x.
Clearly the n-functions x,, x2, ......, X, are real-valued functions 6f n-variables with
domain Rn.The function nj,1 I j I n, is called the j-th projection from Rn to R.
51) For an.y x E R,define f(x) = (x, 0).
Then f : R + R2is a vector-valued function of one-variable.
li) Let D be the open sphere with centre at (0, 0,O) and radius 1 in R3.Then

is a real-valued function of 3 variables with domain D.

vii) For any (x, y) in R2 define g(x, y) = (x, y, 0)
Then, g : R2 + R3 is a vector-valued function of two variables.
viii) For (x, y) E R2define f(x,y) = (excosy, ex siny).
Then f(x,y) is a vector-valued function of two variables with domain R2.
ix) A polynomial in n variables, x,, x2. ........, X, is an expression of the form

Xa kli k2i ,........, kn, x, kil x2 k2i .......x, kni, where kj:s are non-negative integers.
and kli + k2i+ .......+ kni=i.
Advanced Calculus For example, x3y2z+ lox2 yz3 +'8xyz + z5 is a polynomiakin 3 variables, and xy2 + 2xy -
y4 is a polynomial in 2 variables. If we define f : Rn -+ R such that f(xl, x2, ....,....,x,) = a
polynomial in n variables, then f is a real-valued function of n variables.
Remark 3 :As in the case of functions of one variahle, we often define functions of d
'
several variables with the help of a formula. W e n a tunction of several variables is defined
with the help of a formula, then its domain is the set ot all those points where tAi!given
formula is valid. For example, the domain of the function

of three variables is the closed sphere

Similarly, the domain of a function

of two variables is the set R2\ {(O,0)).

Remark 4 :Let f : D 4 Rm,m > 1 be a vector-valued function, where D is a subset of
Rn.Then the function f gives rise to m real-valued functions defined on D, which in tuin
determine f uniquely. The functions are given by (x, of ) (x) = 5 (f (x)), x E D for
1 S j 5 m, where xj denotes the j-th projection,
xj : Rm+ R defined in iv) earlier.
Clearly, f(x) = (x, (f (x)), n2 (f (x)), ...... x, (f (x))).
Conversely, if g,, g2, ...., g, are m real-valued functions defined on D, then these functions
give rise to a unique vector-valued function g on D defined by
g(x) = (g, (x). g2(x). -....,g,(x)).
This means we can break up any vector-valued function into a number of real-valued
functions. As a result, you.wil1 soon see that many times the consideration of vector-valued
functions can be reduced to the consideration of real-valued functions. The functions
g,, ......, g, are usually referred to as the components or component functions of g.

You are already familiar with the graphs of a number of real-valued functions of a real
variable. Now let us see how we can geometrically represent a real-valued function of two
variables.
Mnition 4 : Let f(x, y) be a real-valued function of two variables with domain D. Then
the graph of the function f is the set of points (x, y, z) in the Euclidean space of
3-dimensions such that z = f(x, y), i.e.,

We are giving the graphs of some simple functions here in Fig. 5. ..

Fig. 5 : Graphs of (a) f(x, y) = x -y + 2, (b) f(x, y) = x2 + y2, (c) f(x, Y) = x2 - Y2

 But in most cases you will find that it h not easy to plot the graph of a real-valued function Functions of Several
Variables
of two variables. However, we can visualise the graph with the help of 'level c,wves' defined
below.
Definition 5 :Let f(x, y) be a real-valued funcnon of two variables and let c be a
constant. Then the set of points (x, y) in the plane such that f (x, y) = c is called a level
curve of the function with value c.

I
1
Clearly the level curve f(x, y) = c is nothing but the intersection of the surface z = f(x, y),
i.e., the graph off, with the plane z = c.
Roughly speaking, the graph of a real-value4function of two variables can be obtained by
piling up the level curves f(x, y) = c, as c varies over the range, that is the set of values of
f(x, y). Look at this example.
Example 4 :Let us find the domain and the rar)ge of

ii)
9
+ 16 + z2 = 1, and examine their ievel curves.
i) The domain of the given function is the whole of R2. The range is the set of all real
numbers I 100. The level c w e s are the circles with centre at the origin. See Fig. 6(a)

Fig. 6 : (a) Graph of f(x, y) = 100 - x2 - y2, (b) Level curves

I ii) Note that here we have not expressed the function in the form z = f (x, y), and we cannot
write the value of z explicitly by substituting the values of x and y. Still we can find
the level curves by putting ,+ 2,,
x2
= c. These will be ellipses (see Fig. 7).

I Example 5 : We'll now draw the graph of the function f (x, y) = x

The graph of this function is the plane z = x - y in R3. See Fig. 8.
The level curves are the straight lines x - y = c.
- y.
 Fig. 8 : Graph of z = x -y 1
w e can extend Definition 4 to the case of real-valued functions of three variables.
Definition 6 :Let f(x, y, z) be a real-valued function of three variables with domain D
1
Then the graph of the function f(x, y, z) is the set
Graph off = G(f) = {(x, y, z, w) I w = f (x, y, z). (x, y, z) E D 1 in R4.
Since the graph of a real-valued function of three variables lies in the 4dimensional
Euclidean space, it is not possible to realise it geometrically. We can, however, visualise the
level surfaces defined below.
Definition 7 :Let f(x, y, z) be a real-valued function of three variables and let c be a
constant. Then the set i f points (x, y, z) in space such that f(x, y, z) = c is called the level
surface of the function f with value c.
What are the level surfaces of the function f(x, y, z) = x + 2y + 3z? These are the planes
x + 2y + 32 = c, where c is a constant.
You will agree that the level surfaces of the function f(x, y, z) = x2 + y2 + z2 - a2 are spheres
given by x2 + y2 + z2 - a2 = c, or x2 + y2 + z2 = c + a2, where c > d2- .
Try to do this exercise now.

E 5) Find the domain of the following functions :

1 1
c) f (x, y) = x sin -
X
+ y .sin -Y
1

Just as we can define the sum, product, quotient for functions from R to R,we can define
these algebraic operations onTunctions of several variables too. Let us consider these one by
one.
Ssm of Two Functions :Let f :Dl + R m and g : D2 + Rm,where Dl and D2 are
subsets of Rn.Let D = Dl fl D2 # g. Then the function f + g defined on D b)
(f + g) (x) = f(x) + g(x)
is called the sum of the two vector-valued functions f and g.
Product of Two Functions :Let f : Dl + R and g : D2 + R, where D, and D2 are
subsets of Rn.Let D = Dl n D2 # g .Then the function fg defined on D by
 Functions of Several
Variables
is called the product of the two real-valued functions f and g.
Quotient of Two Functions : Let f and g be the real-valued functions mentioned
above. Suppose that the set
I

i Then the function f/g d e f d on D* by

I is called the quotient of the functions f and g.

I
You must have noticed that we have defined the sum f +,g when f and g are two vector -
valued functions. But we have defined the product fg and the quotient (flg) only for real -
valued functions f and g. This is because, as we have mentioned at the end of Sec. 3.2.2. the
product and quotient of two vectors are not defined.

I We shall now illustrate these operations with some examples.

I Example 6 :i) Let f(x, y) = ysin

D, = Domain off = ((x, y) 1 x # 0 )

1
; 1
and g(x. y) = xsin - - Then
Y

D2=Domainofg= {(x, y) 1 y #0)

I Clearly D , ~ D , = ( ( X , ~I x) # ~ a n dy # O ) # &
i
I
Thus: the sum function
1
(f + g) (x, y) = f(x, y) + g(x, y) = y sin ; + x sin -1 is defined on Dl tl D2 i.e.,
I Y
I R2\{both the axes).
I .
ii) Let f (x, y) = (ex cosy, ex siny) and g (x, y) = (x2, y2) Then the sum function
(f+g) (x, y) = f(x; y) + g(x,y) = (excosy + x2, ex siny + y2)
is gefined on the whole of RZ.
iii) ~ e t f ( x , ~ , z ) = l x( y( 1 2 , g ( x , y , z ) = s i n ( x + y + z ) .
, Then the product function fg is defined by

I and has its domain as R3.

I iv) Let f(x, y) = 2xy, g(x, y) = i 2 + y2.

1 Clearly
D* = ( (x. Y) 1 g(x, y) # O ) = RZ\I(0,O)).
Thus, the quotient o f f and g is defined by

(3 f(aLY1 -
(x9Y, = g(x. y) - x2 + y2
and its domain is Rz\((0. 0 ) .

I Here is an exercise which vou can trv.

E 6) Find the product and the quotient of the following pairs of function. State their
domains in each case.

I
a) f(x,y) = x2 Y, g(x,y) = x2y2
1
b) f(x.y) = sinx + siny, g(x, y) = -cosy, x # 0.
X
Advanced Calculus You have already learnt how to define the composite of two real-valued functions of a real
variable (Sec.6 of Unit 1 in Calculus). To recall the cornnosite of twn fi~nctinnsf. R 4 R
and g :R + R defined by f(x) = x2 and g(x) = sinx, respectively, will be the function
g o f : R + R such that g f (x) = g(f(x)) = sinx2. Here, we extend this notion to,
0

functions of several variables.'

Definition 8 : Let g : D, + Rm,where Dl is a subset of R n and f : D2 + RP, where D2

is a subset of Rm.Suppose that g (D,) c D,.
Then we can define a new function 4 : D,
,

+ RP by setting- q'~ (x) = f(g(x))

- .
Ia
for all x E Dl.
This new function 4 : Dl + RP is called the composite of the functions f and g and is
denoted by f g . If n'= m = p = 1, then this definition coincides with the definition given
0

in the earlier course of Calculus.

Let us See some examples of composite functions.
Example 7 : Let g(x, y) = x2 + xy + y2 be a function from RZ + R and f(t) = sint be a
function from R + R. Then the composite function f g dsfined by 0

(f o g) (x, y) = f(g(x, y)) = f(x2 + xy + y2) = sin(x2 + xy + y2)

is a function from RZ + R.
Note that g 0 f-does not make sense over here.
1
You may also come across functions f and g, for whichf 0 g and g 0 f are both defined, but
are not equal (see E 7a).
Example 8 : Let f(x, y) = (x2 + y2, x+y, xy) be a function from RZ + R3 and
g(x, y, z) = (ex+y,sin (y + z)) be a function from R3 + RZ .
Then the composite function g 0 f defined by

(g O f) (x, Y)= g(f(x, Y)) = g(x2 + y2, x+y, XY)

= (ex' + y2 + + y, sin (x + y + xy))

is a function from RZ + RZ.

You can now easily write the composite of the functions given in the following exercise. i1
- - - - - - - - - - - - - - - I
I

E 7) Find f 0 g and g 0 f, if they exist, for the functions given by

a) f(x, y, 2) = (ex, In (x2+ y2 + I), z2), g(x, Y, Z) = (X + Y, 2y. 52) 1
lo, (x,y)=('%Q) 1

Let us now summarise what we have covered in this unit

3.4 SUMMARY

In this unit we have

1. defined the Cartesian product of sets and discussed the algebraic structure of Rn.
2. introduced a distance function on Rnand defined r-neighbourhood of points in Rn.
3. defined real-valued and vector-valued functions of several variables.
4. inkxluced level curves and level surfaces, respectively, for functions of two and
variables.
5. defined the sum, product, quotient and composite of functions from Rn+ Rm.
3 . 5 SOLUTIONS AND ANSWERS Functions of Several
Variables

El) S 1 : x ~ R n ~ x = ( x 1 ........
, x 2 xn),whereeachxi~R
Now ax = (ax], ax,, ....... ax,). Since axi E R for 1 I i < n, ax E Rn.
S, : Let x = (x,, x,, .....,x,) and Y = (Y,, Y2,....... Y,)

Then a (x + y) = a[ (x,, x,, ........... x,) + (y,, y2, ........yn)l

= [ax, + ay ax,,, + ay,, ..........ax, + ay,]

- (ax,, ax,, ...?....ax,) + (ay,, ay,, ..........., ay,)
-
- a(xl, x2, .........,x,) + a(y,, y,, ......., y,) = ax t ay
S3. S4 follow similarly.
S, : Suppose ax = 0 V x E Rn. Choose x E Rn such that x = (1, 0, 0, .......0).
Then, ax = 0 a(1, 0, 0, ....,0) = (a, 0, 0, ..........0) = 0

Conversely, suppose a = 0. Take x = (x,, x,, ........ x,) E Rn.

Then ax = (ax,, ax,, ...........ax,). But axi = 0 V i = 1, 2, .......,n.
Thus ax = 0.
This is true V x E Rn.
-
E2) e l = (1,0, .......0),e2 = (0, 1 , 0 , .......,01, ....... en = (0, 0, ....... 1)
,Let x = (x,. x,, ........x,) E Rn.
Then
x1 e l + x,e2+ ....... + x,e, = x , (1, 0, .......0) + x 2 ( 0 , 1, ......,0)
+ ...... +X,(O,O ........ 1)
= (x], X, ......, x,)
This shows that every x E Rncan be written as x = x xi ei.
n

i = 1

To prove uniqueness, suppose x can be written as x = C xi el and x = C y, ei where

x , a n d y , ~R V i = 1 , 2......... n.
Then (x,, x,, ......x,) = (y,, y,, ....y,). This shows that xi=yi V i = 1, 2, ......n.
n
Hence x can be written in only one way as x, e,.
I = I

E3) x = e + f = ( 1 , 0 ) + ( 1 , 1)=(2, 1)
y = 2 e + 3 f =2(1,0) +3(1, 1)
= + (393)
= (5,3)
:. , X - y = (2, 1)-(5,3)
'

= (-3, -2)

:. . I x - 1 = 4(-3), + (-2)2
Advanced Calculus
Similarly, ) 2 x - y ) = ) ( I , - l ) ( = f i and(x1 = ( ( 2 , 1 ) ( = 6 .

NOW,(x, y, z) E S *4(x - a12 + (y - bl2 + (Z - c ) ~ < I

3 (x - a)2 + (y - b)2 + (z - c ) <

~ r2 1
(x - a)2 < r2, (y - b)2 < 1-2 and (z -'c32 < r2

*P2c S.
E 5) a) The domain consists of all the points in R2 except those for which x4 + y4 = 0.
Now,x4+)P=O~x4=0andy4=O
-x=Oand y=O
Therefore, domain = R2\ ( (0,O))

b) I ( x , Y ) E R ~ ~ X - ~I =
+OI ( X , Y ) ER ~ ( X +1 Y

d) The open sphere with radius 2 and centre (O,O, 0), since 4 4 - x2 - y2 - z2
has to be positive. '

E 6) a) (fg) (x, y) = f(x, y) g(x, y) = x2y. x2y2= x4y3

Domain = R2

1
b) (fg) (x,y) = - (sinx + siny) cos y
X

x (sinx + sinyJ
COSY

Domain = ((x,y) E R2 I X
y t (2n + 1) 5 n E Z )
7

...e,
I a) ' c 1.1 (x, y, z) = g(f (x, y, z))
\
,I~L.\. In (x2 + y2 + I), z2)
= cc\ + In (x2 + y2 + 1). 21n (x2 + y2 + l), 5z2)
 Functions of Several
Variables

= (eX+y,ln(x2 + 5y2 + 2xy + 1), 25z2)

-
Clearly, f g # g o f, eventhough both are defined.

I
f 0 g does not exist.