Mathematical Methods -

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329
 Chapter Functions Several real variables
15
of
-

several variables because

→ we need
functions of
most
of the
phenomenon nee
study depend on
a
large no
of variables
simultaneously .

→ Ex : demand
of good depends
a on : consumer
tastes income etc
g-price ,
.

15.1 valuables
Functions
of two or more

Definition : [ 151]

variables
function f of
A two
Randy with
domain ☐ is a rule that
assigns specified a

number
fee y) , to each
point @ ) in D. ,
y

Encrypted foe g) ,
= 2kt
Ñy3
f. Edgy ) Azeayb ( constants
bare A
=
a,
,

-
also called Cobb -

douglas function .

what is
fun 2g ) ,
if flag g)
=

Azeayb ?

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

Functions than two variables
of more

variables
-

function f of n se,
,
.
. _

, an with domain
☐ is a rule that
an)
assigns a
specified
vector (x
number

ftd , ,
- . -

s
to each n -

, ,
- - -

,
Rn) in D.

Example :

f- th ,
R2 , - -

g Hn) =
Aiki -1 Azaz -1 - - -
+ an Rn tb

( ai , 92 , - - -

s An and b are constants

this is called a linear

function
-

in n variables .

flag 22 , - -

s Rn ) = A kid Rza ? '

. .
✗ naw

(A) Ai
A
,

>
Az ,
0
)
- -

an are constants

Cobb
this is called a
douglas
-

function

Arithmetic mean
of Ng Nz, pen is
-
-
-

In ( kit Nz -1 + Rn )
= - - -

this is
n variables
also a
function of
.

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

Domains

A domain
of function defineda
formula
bywhich the is the a

largest possible domain in

and unique
value
formula gives meaningful a

Eaaimplei
fee, g) =
Tai +
Ty
since a- I ≥ 0 a ≥1

y≥
0
y ≥ o

So and
all the points at a≥ 1
y
≥ 0
form
the domain
of this
function .

y↑
÷
→

P P P ↑ → ↑ P P
< O
,
1 R→
→

Examples :

JET)
glory )
g÷→→
=
+

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 sexy [ denominator value
'
So -4>0
cannot beano]

x4y2 > 4

and 9-
(ñ+yY≥o
ñ+y2≤ 9

ñ+y2≤ 9
<

domain
canbeshonenbythe following
graph
:

Ñty2=pf
is the equation of
a circle whose
and
if
center is
radius is origin
k .

4sÑ+y2≤ 9
<
→
,ztg\z -
2
3 >a→

-3

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

15.2
Geometric representations of functions of
several variables •

--÷ÑM3ñxy

zP
✗ =
5

se →

Y 20
=

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 2=20

R -110 -1102=100
y

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 -122=100
x2 +
YZ

graph of
functions of
2 variables

z=f(My)
E-
set-up →

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 2--22 +
Cuehen looked
y2 the
from
top ) .

Level curves
for 2
for y )
=

g. the
graph of this
function is
by
cut a
horizontal
plane ( parallel ay planet
to And
resulting
.
the

intersection is projected
That will level
onto ✗
if plane •

the curve
for a
particular
value
of 2 .

So if intersecting plane is z=c

,
the
intersection
projection of the onto
ay
celled level
is
height
at
plane curve a

for f.
a- will consist
of points satisfying the
equation fee g) c. ,
=

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 Examples :
atty
? 8
( Lewd curve with c-
-

8)
Ñty2= 5 ( Level curve with a =
5)
(z=k2+y2 )

2=8 6

←
2=5

Y ,

atty? 8
\
<

% 3g
* →

zz

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

Exampled Suppose FCK , = no
of units producedby a

firm when input of capital is k and

labor is L
that
of .

in
A level curve
for this function is a curve

in K L
-

plane
given by
FUN =
Yo (Yo is a constant)

→ called
Ssoguant ( equal quantity

Any function of
continuity constructed from continuous functions
variables that be n can

by combining the division

operation of addition subtraction , ,

multiplication, and functional composition

wherever it is
is continuous
defined .

Ex :
flag 2) =

sty 8Ñy5
+ a
ay
-182
-

it is a sum
product of powers
ofand .

all
So
firs defined continuous
for xiygz .

Ese :

guy) =Ny-3_
2-4
sexy
This is continuous
for all
points except
when the denominator is zero ◦

set
YZ -4=0
city 2=4) → not
Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi
defined on these
points
9810420329
153 Partial derivatives with two variables

Suppose 2 =
foe g) , .
Let 23 called the partial
,
derivative on with
of Olf a
respect to kg
be the derivative
of foe g) , [Link] x when
is held constant
y .

Also
,
let 23
,
called the partial derivative
of
f OY writ be the derivative
of
y
a or

with respect when xis held

flag)
constant .
toy

Fy ¥
= = rate
of
when
change of
seis
foe g)
,

held constant
[Link] to

yr .

Examples :
fcxiy ) =

a3y sty setup

+ +

8¥ =

3k2g 2kg2 + -11 ( holding y

constant )

23
-12yd -12g ( holding
¥y
= a constant

da flatly fi lay )=2fK

3¥ Fa
= =
-

= -

OR

All these notations mean

the same
thing .

¥y=¥y dy fly lay) filmy 0¥21

= = =
=

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 Higher order partial derivatives

ftriy)
0¥
of and called
2-
,
then
2¥ are

first order partial derivatives

also
Now
2¥ and
¥
are
functions ofay .

if you do partial derivative

again,
Ed:) -3¥ ,
3yP£T¥ya
Eel :&) -3¥ ,
¥1385 3¥
called second order partial derivatives
they
→
are .

Estample : f- trip =

a3y+Ñy2+a+y2
8¥ =

3k2g 2kg2 + -11 ( holding y

constant )

set
2yd -12g ( holding
y
= a constant

1%-2=2*(1-2)
2
Now =

Gay -12g

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 1¥ 3y(¥y)=añ+a

3¥y=¥(¥y)
sat +4
ay
=

322+4
Eyton By (¥a) ay
= =

3¥ 1¥ ( in this )
case
=

y
Approximations to partial derivatives :

false g) ,
≈
flat g) ,
-

flay)
Hy Kip ≈
f- thy-11) -

floe g)
,

The
partial derivative flakey) is
equal to the change in flag ) that
approximately
results
from
increasing
constant
by n one unit while
holding
y
.

'
Similar
interpretation for f- you y )

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 154 Partial derivatives and
Tangent Planes

z=
ftkifo)
the Koryo ) slope ~
Ky
=
=

bitty tray D= slope

\
E-
ftro g)
,

Another
interpretation :

* measures the
steepness in a
particular direction .

So the Lao yo ),
is the steepness in the direction
parallel to positive a- axis .

direction
and
fig tray ) .
is the steepness in the

parallel to positive ants

y
-
.

Planes
Tangent :

The
tangent plane to 2 =

flay) at the
point Kayo RD
with Zo
ftoeoiyo) has the equation:
=

f! laoifo) toe
'
2- Zo No ) 1-
f-atkoifo) G- Yo)
= -

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

Examples :
at
grid the
tangent plane ( Koryo -4=(1/15)
,

Of flay)=Ñ+2xyt2y2

fCbD=5I FILMY )=2Rt2y

fill / 11=4

1-{ lay ) =
ay -12k
, fH41 ) = 6

E9M¥oY==f :(any ) ( . Rko ) -1

fhloloifo)(Y-Yo)
2- 5--4 ( x 1) -16
Ly 1)
- -

5
2- 4k -4
-16g -6
=

4k
-5
-16g -10
z =

2=42
-16g -5

flay )=kH2Mf+2Y2 A :( 41,5 )

-

2--42
-16g -5

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

155 Partial derivatives with variables
many
of 2- -

flank ,
- -
-

Mn )
,
then of /ya is the derivative
,

of f Kimi -

an ) with
respect to ai when all the

other variables Kj ( Iti) are held constant .

There are n
first order partials off .

and
for each
of the n
first order partials off,
we have n second order partials :

Faj(¥ai ) }¥pµ= ziij

=

so in total will have n2 second order

we

partial derivatives .

second order partials

The nxn matrix
of :

) "nn
film five . . .

fink
tzicn ) firm . _ .
"
f- znln)
-

: : :
_

'
em faith . . .

f
NXN
Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329
 This is called Hessian matrix
off
evaluated at a =
Kuku -
-
-

Mn)

" "
and
usually 1-
ij (a) =
f- ji (a) for all igj
So we will have a
symmetric matrix .

(
using Young 's theorem

Estampll :
B , g) B2 AIB
a. (A ,
= 122 -13A -25T -75 -

find the hessian matric

of ACA / Bit)

¥
=3
BI bag -150 B
-1%2,09-9=-25
-
=

82%2=0 ,
Egon .
=

Fsa , %%a=°
Jha
OJAB
-

¥ , 05¥ =

-150-2%-3,32%3=0

Ing -0,8¥,=◦ 3¥
-

_ ◦
,
,

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 So the hessian matrix would be

[ a)
A
AA AAB 9 AT

ABA A BB ABT

A airs

µ :)
=
FBZ 0

1
,
-150-2^8,3
g 0

Young 's theorem

suppose that two Mtn order
partial derivatives of the function
foetida an) involves the same number
,
- -

of differentiations with
respect to each of the
variables , and are both continuous in an

open sets .
Then the two
partial derivatives
necessarily equal at
are all
points in S the .

(
proof beyond
Example :
if m = 2 )
scope

0¥ =
2¥
ON 224 224222

both these
if partials are
continuous .

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 Formal Partial derivatives
definition of :

If 2--404 / Nz Rn)
- -
-

lim grith, )
03 =
f ki ,
- -
-
- -
an -

fkn - -

Hi , Rn)
-

Oki h→ 0
h

If z
ffxiixz
=
,
- -

Nn) has continuous

partial
derivatives order in
of the
first a domain

continuously differentiable
A we call f
,

5- is Ct
in A • We
say
a
Jun -
etion on A •

all partial derivatives to order k

Sf up
exist and are continuous f- is called a
,
Ck
function .

Esiample : Find
firstthe
and second order
partials of function
XYZ sty
2) =3 1- we
wtkiy
-

1508 Quadratic forms in two variables

A quadratic form of two variables is a

function
f- Caryl añ+2bny+cy2 =

where a , b , c are three

real constants .

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 matrix notation
using ,

flay)= K' F) 1×2

(8%2×2%8) 2×1

Hessian
off is
given by
21% b)
The
quadratic form flay)=aÑt2bxytcy2
condition
Type
1. Positive
definite if flay ) >◦ for all
lay) ≠ (0/0)
2. Positive semi
definite if fay) >
◦
for
-

all @
if )

Negative definite if flay ) for

3. <0

all @ iy) -1-(0/0)

4.
Negative semi
definite if flay )≤ 0
for
-

all @if )
5-
tondefinite if flay ) > ◦
for
some values & flay )C0
Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi forsake values
9810420329 .
Example :
imy )=Ñ+y2 , since
atty >0
for
all @ ) ≠ 1010
)
if
it is positive definite
@+ yJ≥o
2
② flay )
=
@ g)+
,
since
for
all @
ip
so it is positive semidefinite
is
y2 negative definite
at
③ flay )= -
-

flag) CITY negative

semidefinite
④ = is
-

⑤ flay ) = at -

y2 is
indefinite

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

The
definiteness of thequadratic form a
strictly
depends on values
of coefficient
a ,b , C •

The quadratic form flay)=aÑ+ zbaytcyis

Positive
definite a >0
,
c >0 &
/ abbot >0

Positive semi
definite a≥ o e ≥o &
-

i :{ I ≥o

Negative definite a < 0

, cog and
1%11>0
Negative semidefinite a ≤0
,
c ≤ o &

1821>-0
sndefinite lab { 1<0
(Recall that lab be /
= are -

b)
Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329
 general quadratic functions in two variables :

foe g) , =

[Link]
This is the
general quadratic function
of Randy .
St can be expressed as
2
foe g) lot E) -12h Get E) ly 2) Cly-1112
,
=
a + +

+d

provided that

§ -12kg =p
2a
And a
£+2b §h+cq2+d=q
bag 2cg q
=
+

So that all the terms match with each other .

your function
Now see that becomes
you can
foe g) ,
= a Lat
)2t2bx*y* -1 Cly92nd
↓
Cat § ) KY -1N
Quadratic form with Linear constraints

suppose Q = art +
2k2g cyf -1
subject to

the linear constraint

patchy = 0 ,
where 91=0 .

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

 pray
So
y
-
=

[Link]
Q =

[Link]
art
= -

Q =

Egg [
art of -

2bpqa4 cpzxt]

Q
# [ [Link]
=

So Q lay ) will be positive (negative

definite subject to the constraint
psetqy
= 0

if Q is positive ( negative ) for all

Cny) ≠
( 0,01
satisfying the constraint
patqy
= 0 .

/ f
° P N
@ of zbpq + copy
- = -

p a b

b a
q

{ So Q
subject
=

artistry eye is positive definite

to constraint
+

} pntqy
= 0

1 : : :| < ◦

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329

Examples :

⑨ Express flay)=2Ñ Yay y2 -32+4 y

-
+

in the
form of
ftdiy) -
alk -1912 -1 (
2b at
g) tf -121 Cly-1nF
+

1- d.

② Examine the the quadratic

definiteness of
form subject to the given linear constraint
at
-2kg -1
YZ subject to
N+y=0

Poonam Kumari, B.A.(H) Economics, SRCC, MSQE, ISI, Delhi 9810420329