D

DIFFERENTIAL CALCULUS
L. Successive differentiation.
Given a function, y=f(r), we denote the first,
, n order differential second,
coefficients of y w.r.t. x
by

Example. If y -3ax +x=0, show that

dy,
da2
20 *=0.
4

Differentiating y° =3ax -x w.r.t. x,

3y y, = 6ax -3* or yy, =2ax -x.
2ax -x
d'. y (2a -2r)-(2ax -x)(2yy,)
de2
Multiplying the nr. and dr. by yand pulling out 2 as the
common factor fronm the nr.

,y' (a -x) - (2ax-) ('y) (1)

But y' =3ax -x and y'y,= 2ax - *". Substituting in (1),

, (3ax' - ' a-x)-(2ax-2ax - )
y°
=2 (3a' x-3ax' -ar +]- (4a? x- 4ax +x]

=2: (0) +*°0) +*(-a 2a?

2
DIFFERENTIAL CALCULUS

2a2
=0.

Derivatives w.r.t x when * =ft), y = g().

In this case,

dydx =y1 dy/dt

dx/dt

d'y dy,/dt
dx da/dt
For example, if
K=a cos t, y=b sint,
then

dy/dt bb cos t b
Y1= dx/dt cot t,
-a sint

dy, /dt (-b/a) (- cosec t) cosee t.

y2 dx/dt -a sin t?

EXERCISES
h'- ab
1. If a+ 2hxy + by = 1, then show that y, = (hx + by)
[Hint: Differentiating both sides w.r.t. x and dividing by 2,
a(23) +2h(«y, +y) +b(2y9y) =0 or y1 (hx +by) = -(ax +hy).
ax +hy
Y17 hx +by
(hx +by)(a +hy,) - (ax + hy)(h +by,)
(hx +by)'
y,(h- ab) x- (h- ably (hx -ab
+byj?-yt)
(hx +by)²
h-ab a' ++2hxy +by-ab,
(hx + by) hx + by (hx +by)"
SUCCESSIVE DIFFERENTIATION 3

20' xy
2. If +y =3axy, show that y, =(ar-y
IHint: Differentiating ° +y° =3axy w.r.t. x and dividing by
3.
y, (ax -y=-ay.
Again diferentiating w.r.t. x,
y, (ax-y +y, (a - 2yy,) =2r -ay,
i.e., y, (ar-y) =2 k+yf - ay,)
2

-2 **yay ax-y
2x(ax -y+yo -ay-a(g' -ay) (ar-y)
(as-)
2la' (6y) +a 0) +a(-3y+sy(+)
(ax-y
collecting coefficients of a, a, ain the numerator,
= 2(a'xy +a(-3r+sy (3axy)] 2n xy
(ax-y'? (ax-y
Dividing by ax -y', we get the value of y, as stated.]
3. If x =a(t- sin t), y =a (1+ cos t), find y, in t.
dy/dt -a sint - 2a sin,42 cos 2
[Hint: y1 =- cot 2.
dx/dt a (1- cos t) 2a sin
dy,/dt ½ cosec 2 cosec 2l
dx/dt 2a sin t 4a

1.1.1. n"h derivatives.

We shall obtain the nh derivatives of
(i) (ii) (ax +b)" (iii) log (ax +b)
JACOBIAN
31

1.2/ Jacobian.
If u andvare two functions in x and y,
the determinant

is called the Jacobian of u, v w.r.t x,

y. This Jacobian is
denoted by
du, v) Or
J(u, v; x, ).
d(x, y)
This definition can be extended to three
u, U, w in x, y, z as functions

du, v, w) uy
d(x, y, 2) Vy
W, Wy W,
Properties of Jacobians. Two important properties are
embedded in the following bookworks. The proof is given for
two variables only though the result is true for any number
of variables.

Bookwork 1. Show that JJ' = 1, that is,

du, v) d(z, y) = 1.
d(x, y) d(u, v)
Here u and v are independent of each other and are
functions of the independent variables x and y and vice
versa. Therefore
du
du
du
or , , tu, y,=0,
du
du
du
du or , , tu, y, = 1.
d(u, v) d(x, y) uy
dr, y) du, v) Uy Yu
 DIFFERENTIAL CALCULUS

u, U and u, v an.
Bookwork 2. If U,Vare functions ofthat
themselves functions of x, y, then show
aU, V) U, V) du,v)
dr, y) O(u, v) (,y)"
Now it follows that
U,-U, u, +U, U,,
U, =U, u, +U, Vy:
V,=V, u, +V, U,
V,=V, u, +V, y
a(U, V) d(u, )
d(u, v) Ox, y) V, V,
U, u, +U, y

U, U, aU, V)
V (x, y)
EXAMPLES

Example 1. If x =rcos 0, y =r sin 0, then find

d(x, y) (2004)
dr, 0)
cos - r sin 0
d(r, 0) sin 9 r cos

=r (cos e+ sin 0) =r.

dr, 0)
Example 2. If x =r cos 0, y =rsin 0, find dx, y)' (2003)
We know that

x, y) d(r, 9)-1.
d(r, e) Ox, y)
JACOBIAN
33

But, from the previous example, Hence

dr, 0)
r Or, 9) =1. d(r, 0) 1
d(x, y)
d(x, y) r
Note. This result can also be obtained by
I,, r O 8, from r =*+y', tan =y and using calculating

Example 3. f x=rcos O, y=r sin 0, 2=2, (where

r.6, z are cylindrical coordinates), show that
d(x, y,Z) =r.
dr, 6, z)
We have x,= cos , o =-rsin 0, x,=0,
y,= sin 0, Ye =r cos 0, Y,=0,
2,=0, Z0 =0, 2,=1.

d(x, y, 2) cos -r sin 0

sin 9 r cos 0 0=r.
dlr, 0, z) 0 1
Example 4. If x =r sin cos ), y=rsin 0 sin ),
z =r cos 0, find the Jacobian of x, y, z w.r.t r, 0, 0.
(1990, 91, 00)

J=|Y,
N

sin cos r cos e cos ) -r sin 0 sin )

= sin sin ) r cos sin r sin cos
COs -r sin
Taking out r from Co and rsin 0from Ca, we have
sin 0 cos cos cos - sin o
J=r'sin 0 sin 0 sin cos sin COs )
cos - sin 0

Multiplying R, by cos 0, we have

34 DIFFERENTIAL CALCULUs

cos 0 cos - sin cos

0 cos
Jsin 0 sin cos 0 sin )
coe sin sin - sin 0 0
cos
sin 9 cos 0
sin 9 cos 0R, ’ R, + sin R.
sin 0 sin¢ cos sin 0
- sin 9
cos
Psin (sin 0 (sin 0cos ) - cos 0 (- cos cos ))
0s
Psin (cos ] =r sin 9.
cos
Example 5. If u=+y+2, v=x ty tz and
that
w=1y+ yz +zx, then show
du, u, w) -0. (1991)
d(r, y, 2)
20c 2y 22
Uy 1
J= du, u, w) Vy
1 1
d(x, y, z) |ytz zt* +y
Wy Wz
2x 2y 22
= 1 1 1
I+y+z +y tz x +y +z
2x 2y 22
=0.
=(z ty +z) 11 1
1
1
1

EXERCISES

1. If u=x+y, y=uv, then show that = u,

Ou, v)
Hint: Solving for x, y, we get x=U- uV, y = uw]
2. If u=-y', v=xy, then show that
1
O(u, v) 2?+y
Hint: au, v) 2x-2y = 2(¢ +y). But we know that
d(, y)
d(u, v) dE, y) = 1]
O(,y) (u, v)
JACOBIAN
35

8. If u=x-y', u=+y,show that d,3)1

du, v) &xy'
(2002)
4. If u =x(1- y), V=Xy, show that JJ' =
1.
(Hint: J=
O(*, y) Now x=u+u, and
1
So J =
"yu+U
(u +u (u +u)'
5. If x=a cosh ucos u, y =a
sinh usin u, show that
a
l(u, v) 2 (cosh 2u - cos 2u).
(Hint: d(z, y) a sinhu cos v -a cosh usin U
d(u, v) a cosh u sin v a sinh u cos U Use

cos? u= t cOS
2
20 sin-1- cos 2u
2

6. If u = 2 v=, w-, then show that

(u, U, W)=4.
dx, y, 2)
(Hint: In J multiply R,, R,, R, by x,y,2 and take out
the common factors yz, zx, xy from C, Cz, C
7. If =x+y t2, uw =y +2, uvw =2, then show that

du, u, w) 1 (2004)
Ox, y,z) u'v

(Hint: x =u - Ww, y= uw- www, z = uvw. Use the formula

(u, U, w)-1
Ox, y, z) du, v, w)'
36 DIFFERENTIAL CALCULUS

1.3 Maxima and minima (Several variables)

We shall consider the maxima and minima of
of several variables. functions
Maxima and minima of functions of two
variables.
Suppose f=fx, y) is a continuous function in xand y.
Consider the values of fx, y) for various values of x and
i.e., the values of fx, y) at various points (I, y). If ther
exists some circle having its centre at (a, b) such that the
values of fx, y) at all the points inside the circle is less th2
its value at (a, b), then fx, y) is said to have a maximum
at (a, b). This maximum value is fa, b).
If the values of fx, y) at all the points are greater than
the value of fx, y) at (a, b), then fx, y) is said to have a
minimum value at (a, b). This minimum value. is fa, b).
Using Taylor's expansion for fr, y), namely,
fa +h,y +k) -fz,v) =(hf, +kf)
+

the points at which fx, y) has a maximum or a minimum

value, if it exists, can be obtained.
The necessary,condition for f, y) to have a maximum
or a minimum at (a, b) is that, at (a, b), both of f, and f
should vanish. That is, at (a, b)
f.=0, 6,= 0. (1)
We cannot say that, at all such points, f, y) have extreme
values. If, in addition to f =0 and f, =0, we have at (a,b),
fefy> fy ie., fufy - fy is +ve, (2)
then fx, y) has an extreme value at (a, b).
Note. It is noteworthy that, if (2) is satisfed, then both
of f and f are of same sign, that is, either both are
positive or both are negative.
MAXIMA AND MINIMA 37

Now the sufficient conditions for the existence of the

extreme value is that, besides (1) and (2)
) if f (or f) is -ve, then fx, y) has a max. value at
(a,b) and it is fla, b) and
(ii) if f (or f) is +ve, then fx, y) has a min.
value at
(a, b) and it is fa, b).
Stationary point or critical point.
a function f, y) has an extremum, is The point at which
called a stationary
point or a critical point of the function,.

Working Rule
Step 1. Find f, and f
Step 2. Solve the simultaneous equations f.=0 and
f=0 for x and y.
Step 3. Find the points at which both f,
and f, are zero.
Step 4. Find f,ffr and choose the points, from
step 3, at which
fxfy-fy is +ve.
(If this condition is satisfied, then both
f and f will be of same sign.)
Step 5. Suppose (a, b) is a chosen point. Then,
(i) if f (or f) is ve at (a, b), then fr, y) is
a max. at (a, b). "This max. is f(a, b).
(i) if f (or fy) is tve at (a, b), then fr, y) is
a min. at (a, b). This min. is f(a, b).
Situations at other points (a, b)
)IEf-f, =0, fsy- f is -ve at (a, b), then fx, y)
is neither a max. nor a min. at (a, b). Here (a, b) is called
a saddle point.
(i) If f, =f,=fafn-f=0, then further consideration
is, needed to arrive at a conclusion.
38
DIFFERENTIAL CALCULUS

EXAMPLES

Example 1. Find the minimum value of

+y'-4r - 2y +10.
The expression can be rewritten as sumn of Squares as
(x-2)° +( - 1)° +5.
It is a minimum when the first two become when Zeros.
i.e.,
x=2, y= 1. Thereby the minimum is 5.
Example 2. Find the max. and min. values of the
function fr, y) =x'y --y'. (1994, 95, 96)
f,- 2ry' - 2x =2x(y- 1),
f= 2ry- 2y =2y(*" -1).
f, =0 gives x(y- 1) =0. .. x=0, y=1, y=-1. (1)
f,=0 gives y(r"- 1) =0. y=0, x = 1, x=-1. (2)

To find the points at which f, =0 and f, = 0 take one

x-value from (1) and one y-value from (2). Then take one
x-value from (2) and one y-value from (1). Thus we have the
seven points
(0, 0), (1, 1), (1, -1), (- 1, 1), (-1, -1)
at which both f, and f, are zero. Now
f=2y -2, fy =2r" -2, f =4ry.
f y-f =4°- 1)-1) - 16x'y' =A, say.
At (0, 0), A +ve. So flx, y) is an extremum there. But
fx is -ve there because f(0, 0) = -2. Therefore fx, y) is a
max. there. This max. is fO, 0) or 0.
At all the other 4 poinis A is -ve. Hence these points
are saddle points.
Note. We cannot get the point taking xand y values
from (1) only because the point will not satisfy f, = 0.
Similarly we cannot get the point taking x and y values
from (2) only.
MAXIMA AND MINIMA
39

Example 3. Find the max. or min. values of the

funetion fx, y)= 2(**-y)-+y. (1993, 94, 96)
f.= 4x-4r" =4r(1-x),
f=-4y t 4y'= - 4y(1-y).
)f=0gives x(**- 1) = 0. .. x=0, 1, -1.
() ,=0 gives y(y" - 1) =0. .. y=0, 1, - 1.
So there are 9 points at which both f. and f. are zero. They
are

(0, 0), (0, 1), 0, - 1),

(1, 0), (1, 1), (1, - 1),
(-1,0), (-1,1), (-1, - 1).
f=4-12x", f =-4 +12y', f =0.
fx fy- fy'=(4- 12*-4+ 12y3- o²
= - 16(1 -3*)(1 -3y); say A.
1. At (0, 0), A is -ve. So (0, 0) is a saddle point.
2. At (0, 1), Ais fve. And f is +ve. So fx, y) is a min.
This min. is - 1,
3. At (0, - 1), Ais +ve. And f is +ve. So fx, y) is a min.
This min. is - 1.
4. At (1, 0), A is +ve. And f is -ve. So fa, y) is a max.
This max. is 1.

5. At (1,1), A is -ve. So (1, 1) is a saddle point.

6. At (1, - 1), A is -ve. So (1, - 1) is a saddle point.
7. At (- 1,0), A is +ve. And f is -ve. So f , is a max.
This max. is 1.
8. At (- 1, 1), Ais -ve. So (-1, 1) is a saddle point.
9. At (- 1, -1), A is -ve. So (- 1, - 1) is a saddle point.
Remark. The above 9 results can be grouped and
expressed in few words.
 DIFFERENTIAL CALCULUS

Example 4. Determine the points where the function

'yty - ary has a maximum or a minimum. (2003)
f(«. y) =xy txy'- axy
f= 2xy +y' - ay, f, =+ 2ry -ax
y=0 (1)
f=0, when 2r +y-a =0 (2)
x=0 (3)
f,=0, when I+ 2y-a =0 (4)
are given by
The points at which both f and f, are 0
and (3) as (0, 0),
(1)
(1) and (4) as (a, 0),
(2) and (3) as (0, a),
(2) and (4) as (43, 43).
Also we have

f-2y, fy =2x, f=2r +2y -a

ff-fg= 4y -(2r +2y - a)' =A, say.
At (0, 0), (a, 0), (0, a), A is - ve. So these three points
are saddle points. However, at (43, 3)
4a 2a, 2a - a
A= + ve.
9 3

fx=2xg' + ve.

These two re: lts imply that f(r, y) is a minimum at

(a3, a3).
Example ). Find the max. and min. values of the
function

fa,y) =xy +41 (1995, 04)

1 1
f=y - f,-x
Therefore f. =0 and f, =0 when
 MAXIMA AND MINIMA 41

x'y =1, ry =1. (1)

To solve (1), we substitute y= in xy² = 1Thus,
x=1 or x=1 or x= 1.
From xy =1, y=1. So we get (1, 1) and only at (1, 1),
f.=0 and f =0. Now
2

4
1, say A.
At (1, 1), A is +ve and f. is +ve. Hence fx,y) is a min. at
(1, 1). This min. is 3.
Example 6. Find the points on the surface 2=xy +1
nearest to the origin.
Let (x, y, 2) be a point on the surface. Then the square
of its distance f from the origin is r+ y' +2 where
z=ry + 1. So f becomes a function of two.variables
f=r +y'+xy +1.
f-2x +y, f,= 2y +x, f=2, fyy = 2, fy =1.
ffy-fy =2x2-l =3>0; fg >0.
So f is a minimum for the values of x and y given by
f.=0 and f, =0 or
2x +y= 0, 2y +x=0. or x=0, y=0.
The z coordinate of the respective points on 2=xy + 1 are
obtained by setting x =0, y =0 in z =xy + 1 as
2=1 or 2=t1.
Thus the points of z=xy +1 nearest to the origin are
(0, 0, 1), (0, 0, -1).
EXERCISES
1. Find the maximum and minimum values of
fu, y) = ?+y+42
x y
4
(Hint: f, =0 =x=1,f=0 y-l. A- 2|2 is

+ve at (1, 1). Minimum = 6]

4
 DIFFERENTIAL CALCULUS

maximum values of
2. Find the minimum and
f(x. y) = xy + y
At (a, a), A =4- 1,
(Hint: f=0.f,=0 =a, y =a;
Minimum 3a at (a, a)l of
Find the maximum and minimum values
3.
fx, y) =x+y'-3axy. -9a'. At
give (0, 0), (a, a). A= 36xy
(Hint: f,-0,f, =0 saddle point. At (a, a), A is
(0, 0), Ais -ve. So (0, 0) is a
= -a]
+ve and fr,y) is a minimum minimum values of
4. Find the maximum and
(1997, 99)
fr, y) =x+y° -3ary.
(Hint: In the previous sum set a =1]
values of
5. Find the maximum and minimum
fx, y) =x+y'+x+y +xy. (1997)
(Hint: , = 0, f, =0 give (- 3, - 3). Here fr, y) is a min.

1.3.2. Lagrange's method of multipliers.

To discuss this method we shall restrict to a function of
3 variables. Suppose we want to find the extreme values of
fa, y, z) where x, y, Z are constrained by the relation
O(, y, z) = 0. (1)
Now
df=f, dx +f, dy +f, de, (2)
0= 0, dx +0, dy +o, dz. (3)
Multiplying (3) by à and adding to (2).

At an extremum point, since df= 0 and since dx, dy, dz

can be arbitrary, the coefficients of dx, dy, dz are zero.
Therefore
f +A0, =0
I,t Ao, =0 (4)
f, +A0, =0