Module -3 Jacobians , Taylor's Series, Maxima and Minima

Jacobians
Changing variable is something we come across very often in Integration. There are many
reasons for changing variables but the main reason for changing variables is to convert the
integrand into something simpler and also to transform the region into another region which is
easy to work with. When we convert into a new set of variables it is not always easy to find the
limits. So, before we move into changing variables with multiple integrals we first need to see
how the region may change with a change of variables. In order to change variables in an
integration we will need the Jacobian of the transformation.

If f1, f2, . . . ., fn are n differentiable functions of n variables x1, x2, . . ., xn, then the determinant

f 1 f 1 f 1
. .
x1 x 2 x n
f 2 f 2 f 2
. .
x1 x 2 x n
. . . . .
. . . . .
f n f n f n
. .
x1 x 2 x n
is defined as the Jacobian of f1, f2, . . . ., fn with respect to the n variables x1, x2, . . ., xn and is
 ( f 1 , f 2 ,..., f n )
denoted by .
 ( x1 , x 2 ,..., x n )

 u u 
 (u, v)  x y 
If u and v are functions of x and y, then J(u,v)=  
 ( x, y )  v v 
 x y 
 u u u
x y z
 (u, v, w) v v v
If u, v and w are functions of x, y and z, then J(u,v,w)= 
 ( x, y , z ) x y z
w w w
x y z

Properties of the Jacobian

1. Chain Rule for Jacobians: If u and v are functions of independent variables r and s and
each of r and s are functions of the variables x and y, then u and v are functions of x and
 (u , v)  (u , v)  (r , s )
y. Further the jacobians satisfy the chain rule  .
 ( x, y )  ( r , s )  ( x, y )
2. If u and v are functions of x and y, then x and y can be solved in terms of u and v. Then
 (u , v)  ( x, y )
.  1.
 ( x, y )  (u , v)
3. If u, v and w are functions of x, y and z and if u, v, w are functionally related or
 (u , v, w)
dependent then 0
 ( x, y , z )

Problems

1. If u = xyz, v = x2 +y2 + z2, w = x + y + z, find J (x, y, z)

 ( x, y , z ) 1
Solution J (x, y, z) = 
 (u, v, w)  (u, v, w)
 ( x, y , z )

u u u
x y z yz xz xy
 (u, v, w) v v v
  2 x 2 y 2 z  2( x  y )( y  z )( z  x)
 ( x, y, z ) x y z
w w w 1 1 1
x y z

 ( x, y , z ) 1
Therefore 
 (u , v, w) 2( x  y )( y  z )( z  x)

 (u , v, w)
2. If u = x + y + z, u2v = y + z, u3 w = z, show that  u 5
 ( x, y , z )
Solution: ux = uy = uz = 1
 vx =
 2 ( y  z ) , v u  2( y  z ) , v u  2( y  z )
y= z=
u3 u3 u3
 3z  3z 1 3z
wx = 4
, wy= 4 , wz = 3  4
u u u u

u u u
x y z 1 1 1
 (u , v, w) v v v  2( y  z ) u  2( y  z ) u  2( y  z )
 
 ( x, y, z ) x y z u3 u3 u3
w w w  3z  3z u  3z
x y z u4 u4 u4

1 0 0
 2( y  z ) 1 1
 c 2  c 2  c1 , c3  c3  c1
u3 u2 u2
 3z 1
0
u4 u3
1
 5
u

x y  (u , v)
3. If u  and v  tan 1 x  tan 1 y, show that 0
1  xy  ( x, y )
Solution: Let x = tanθ and y = tanφ,
tan   tan 
then u =  tan(   )
1  tan  tan 
And v = θ + φ

 (u , v)  (u , v)  ( ,  )
 .
 ( x, y )  ( ,  )  ( x, y )
uθ = sec2(θ + φ) , uφ = sec2(θ + φ), vθ = vφ = 1

 (u, v) sec 2 (   ) sec 2 (   )

 0
 ( ,  ) 1 1
 (u , v)
Thus  0.
 ( x, y )
4. Show that u = x3 – y3z3, v = x2 + y2z2 + xyz and w = x –yz are functionally dependent and
also find the relation.
 (u, v, w )
Solution: If u, v, w are functionally dependent then 0
 ( x, y, z)
3x 2  3y 2 z 3  3y 3 z 2
 (u, v, w )
 2 x  yz 2 yz 2  xz 2 y 2 z  xy
 ( x, y, z)
1 z y
3x 2  3y 2 z 2  3y 2 z 2
 yz 2 x  yz 2 yz  x 2 yz  x taking z common from c 2 and y from c 3
1 1 1
 0 ( Two columns are identical )
Since x  y 3 z 3  ( x  yz)( x 2  xyz  y 2 z 2 )
3

u = v w is the relation between the three variables.

5. If u = excosy , v = exsiny, where x = lr + sm, y = mr – sl, verify if

 (u , v)  (u , v)  ( x, y )
 .
 ( r , s )  ( x, y )  ( r , s )
6.  (u, v) e x cos y  e x sin y
 x  e2x
 ( x, y ) e sin y e cos y
x

 ( x, y ) l m
  l 2  m 2
 (r , s) m  l
u = elr+smcos(mr-sl), v = elr+smsin(mr-sl)
ur = lelr+smcos(mr-sl) - m elr+smsin(mr-sl)
us = melr+smcos(mr-sl) + l elr+smsin(mr-sl)
vr = lelr+smsin(mr-sl) + m elr+smcos(mr-sl)
vs = melr+smsin(mr-sl) - l elr+smcos(mr-sl)

 (u, v) u r us  (u, v)  ( x, y )
  e 2 x (l 2  m 2 )  .
 (r , s) v r vs  ( x, y )  ( r , s )
 Useful Links for this topic

1. http://mathwiki.ucdavis.edu/Calculus/Vector_Calculus/Multiple_Integrals/Jacobians
2. http://www-astro.physics.ox.ac.uk/~sr/lectures/multiples/Lecture5reallynew.pdf
3. http://math.oregonstate.edu/home/programs/undergrad/CalculusQuestStudyGuides/vcal
c/jacpol/jacpol.html
4. http://www.google.co.in/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CCMQFjA
B&url=http%3A%2F%2Fwww.tcc.edu%2FVML%2FMth163%2Fdocuments%2FJacobian
s.pptx&ei=1DKSVZfbG867uAS_t4CABg&usg=AFQjCNHQDmFpTK-
pU16sC61WTkwouEvUFA&bvm=bv.96783405,d.c2E
5. http://math.etsu.edu/multicalc/prealpha/Chap3/Chap3-3/printversion.pdf

Taylor's Series
Statement : Let be a function of two variables which possess continuous partial
derivatives at all points Then

Another form of Taylor series :

Maclaurin's series :

The Taylor series expansion of about the point (0, 0) is called Maclaurin's series.
+ ...

Problems

1. Find the Taylor series expansion of upto second degree terms.

Solution:

Taylor series expansion of upto second degree term is given by

Let

Taylor series expansion of upto second degree terms

i.e.,

2. Expand in powers of and upto and including second degree term.

Solution:

Taylor series expansion of about the point i.e., in powers of and

upto second degree term is given by
Let

Taylor series expansion of in powers of and upto second degree

term is given by

i.e.,

i.e.,

3. Expand as a Taylor series in the neighbourhood of (1, 1) upto second degree term

Solution:

Taylor series expansion of in the neighbourhood of upto second degree term is

given by

Let
Taylor series expansion of in the neighbourhood of (1, 1) upto second degree
term is given by

i.e,

i.e,

4. Using Taylor series show that

Solution:

Taylor series expansion of upto third degree term is given by

Let
Taylor series expansion of upto third degree term is given by

i.e.,

i.e.,

5. Expand in powers of upto fourth degree terms.

Solution:

Taylor series expansion of upto third degree term is given by

Let
Taylor series expansion of in powers of upto fourth degree terms

i.e.,

Maxima and Minima of functions of two variables

The problem of determining the maximum or minimum of a function is encountered in

geometry, mechanics, physics, and other fields, and was one of the motivating factors in
the development of the calculus in the seventeenth century.

A function of two variables can be written in the form z = f(x, y). A critical point is a point
z z
(a, b) such that the two partial derivatives and are zero at the point (a, b). A
x y
relative maximum or a relative minimum occurs at a critical point.

A critical point is a maximum if the value of f at that point is greater than its value at all its
sufficiently close neighboring points.
A critical point is a minimum if the value of f at that point is less than its value at all its
sufficiently close neighboring points.

A critical point is a saddle point if the value of f at that point is greater than its value at
some neighboring point and if the value of f at that point is less than its value at some
other neighboring point. Saddle point is a point which is neither a maximum nor a
minimum.

Working rule for identifying critical points of the function z = f(x,y) and to classify
them

z z z z
Step 1: Find the partial derivatives and . Solving  0 and  0 gives the
x y x y
critical points (a,b) at which a maxima or minima may exist.

2z 2z 2z

Step 2: Find the value of r  , s  and t  at all the points (a,b) got in
x 2 xy y 2
step 1.

Step 3:
i. If r < 0 and rt – s2 > 0 the f(x,y) has a maximum point at (a,b) and the
corresponding maximum value is f(a,b).
ii. If r > 0 and rt – s2 > 0 the f(x,y) has a minimum point at (a,b) and the
corresponding minimum value is f(a,b).
iii. If rt – s2 < 0 the f(x,y) has neither a maximum nor a minimum point at (a,b)
and the point is called a saddle point.
iv. If rt – s2 = 0 the further investigation is required to classify the point

Problems
1. Find the maxima and minima of the function, if any, for the function
f(x,y) = y2 + 4xy +3x2 +x3

Solution: fx = 4y + 6x + 3x2, fy = 2y + 4x. Equate fx and fy to zero

4y + 6x + 3x2= 0 ……(1)
2y + 4x = 0 ……(2)
Solving equations (1) and (2) we get the critical points (0,0) and (2/3,-4/3)
r = fxx = 6 + 6x, t = fyy = 2, s = fxy = 4
 Critical Point r rt – s2 Classification
(0,0) 6 -4 Saddle point

(2/3,-4/3) 10 4 Minimum point

The point (2/3, -4/3) is a minimum point of the function and the minimum value
F(2/3,-4/3) = -4/27

2. Find the maxima and minima of the function f(x,y) = xy (a – x - y)

Solution: fx = ay – 2xy –y2, fy = ax – x2 – 2xy. Equate fx and fy to zero

y(a – 2x –y) = 0 ……(1)

x(a – x –2y) = 0 ……(2)
Solving equations (1) and (2) we get the critical points (0,0), (a,0), (0,a) and
(a/3, a/3).

r = fxx = - 2y, t = fyy = - 2x, s = fxy = a - 2x - 2y

Critical Point r rt – s2 Classification

2
(0,0) 0 -a Saddle point

(a,0) 0 -a2 Saddle point

(0,a) -2a -a2 Saddle point

(a/3, a/3) -2a/3 a2/3 Maximum or

minimum point

(a/3, a/3) is the only point which could be either be a maximum or a minimum.
r depends on the value of ‘a’.
r = -2a/3 < 0 if ‘a’ is positive
r = -2a/3 > 0 if ‘a’ is negative
Hence f(x,y) has a maximum at (a/3, a/3) if ‘a’ is positive and has a minimum at
(a/3, a/3) if ‘a’ is negative.

a3
The value is f(a/3, a/3) =
27
3. Examine the function f(x,y) = x3 + 3xy2 -15x2 +72x -15y2 for extreme values

Solution: fx = 3x2 +3y2 -30x +72, fy = 6xy – 30y. Equate fx and fy to zero
 3x2 +3y2 -30x +72 = 0 ……(1)
6xy – 30y = 0 ……(2)
Solving equations (1) and (2) we get the critical points (5,1), (5,-1), (4,0) and
(6,0).

r = fxx = 6x-30, t = fyy = 6x-30, s = fxy = 6y

Critical Point r rt – s2 Classification

(5,1) 0 -36 Saddle point

(5,-1) 0 -36 Saddle point

(4,0) -6 36 Maximum point

(6, 0) 6 36 Minimum point

(4, 0) is a maximum point and the maximum value is f(4,0) = 112

(6, 0) is a minimum point and the minimum value is f(6,0) = 108

4. Find the extreme values of the function u  x 2 y 2  5 x 2  8 xy  5 y 2

Solution: ux = 2xy2 -10x - 8y, uy = 2x2y – 8x -10y. Equate fx and fy to zero

2xy2 -10x - 8y = 0 ……(1)

2x2y – 8x -10y = 0 ……(2)
Since (0,0) satisfies both (1) and (2), (0, 0) is a critical point. To get the other
points rewrite equation (1)

8y
From (1) we get x  …….(3).
2 y 2  10
Substitute this in (2)
2
 8y   8y 
2 2  y  8 2   10 y  0 …….(4)
 2 y  10   2 y  10 
Solving (4) we get y = 3, -3, 1, -1 Substitute these values in (3) we get the critical
points (0, 0), (1,-1), (-1,1), (3,3), (-3,-3).

r = fxx = 2y2 -10, t = fyy = 2x2 -10, s = fxy = 4xy - 8

 Critical Point r rt – s2 Classification
(0,0) -10 36 Maximum point

(1,-1) -8 -80 Saddle point

(-1,1) -8 -80 Saddle point

(3, 3) 8 -720 Saddle point

(-3, -3) 8 -720 Saddle point

(0,0) is a maximum point and the maximum value is f(0,0) = 0.

5. Show that x = a/2, y = a/3 makes the function u = ax3 y2 – x4 y2 – x3y3 a

maximum.

Solution:
u x  3ax 2 y 2  4 x 3 y 2  3 x 2 y 3 .........1
u y  2ax 3 y  2 x 4 y  3 x 3 y 2 .............2 
Put x = a/2 and y =a/3 in both equations (1) and (2)
Since both ux and uy are zero at ( a/2,a/3), it is a critical point.
u xx  6axy 2  12 x 2 y 2  6 xy 3
u yy  2ax 3  2 x 4  6 x 3 y
u xy  6ax 2 y  8 x 3 y  9 x 2 y 2

a4
r at (a/2, a/3) =  which is negative for any value of ‘a’.
9
a4
t at (a/2, a/3) = 
8
a4
s at (a/2, a/3) = 
12
2 a8
rt – s at (a/2, a/3) = which is positive for any value of ‘a’.
144
Since r is negative and rt – s2 is positive the point (a/2, a/3) is a maximum point
 Useful Links for this topic

1. http://personal.maths.surrey.ac.uk/st/S.Zelik/teach/calculus/max_min_2var.pdf
2. http://www.maths.manchester.ac.uk/~mheil/Lectures/2M1/Material/Chapter2.pdf
3. http://tutorial.math.lamar.edu/Classes/CalcIII/RelativeExtrema.aspx
4. http://www.ccs.neu.edu/home/lieber/courses/algorithms/cs4800/f10/lectures/11.4.
Maximizing.pdf
5. http://www.maths.manchester.ac.uk/~ngray/MATH19662/Section%204%20-
%20Functions%20of%20Two%20Variables.pdf

Constrained Maxima and Minima

Sometimes we may require to find the extreme values of a function of three ( or

more ) variables say which are not independent, but are connected by a
relation say . The extreme values of a function in such a situation is called
constrained extreme values.
In such situations, we use to eliminate one of the variables, say z
from the given function, thus converting the function of three variables as a function of
only two variables. Then we find the unconstrained maxima and minima of the converted
function.
When this procedure cannot be used, we use Lagrange's method.

Lagrange's Multiplier Method

Sometimes we may require to find the maximum and minimum values of a
function where subject to the constraint .
(1)
We define a function
where λ is the Lagrange's multiplier independent of x, y, z.
The neccessary condition for a maximum or minimum are

(2)

(3)

(4)
Solving the four equations (1), (2), (3) and (4) we get the values of x, y, z, λ which give
the extreme values of

Problems
1. Prove that the stationary values of where occur at

Solution:

Let

and =

implies ⇒
(1)
implies ⇒
(2)
implies ⇒
(3)

From (1), (2) and (3) we get

i.e.,
i.e., = = = =

consider = ⇒

consider = ⇒

consider = ⇒

Thus is stationary at this point .

2. Find three positive constants such that their sum is a constant and their product is
maximum.
Solution:
Let the three positive constants be such that .
Let
and (1)
We have to maximize subject to the constraint
Let =
 implies ⇒
(2)
implies ⇒x

(3)
implies ⇒
(4)

From (2), (3) and (4) we get

Consider ⇒
Consider ⇒
Therefore
Substituting in (1), we get ⇒ ⇒
Therefore , .
Hence the three numbers are .

3. Split 24 into three parts such that continued product of the first , square of the second
and cube of the third may be minimum.
Solution:
Let the three parts be such that .
Let (1)
We have to minimize subject to the constraint
Let =
implies ⇒
(2)
implies ⇒

(3)
implies ⇒
(4)

From (2), (3) and (4) we get

Consider ⇒
Consider ⇒
Substituting in (1), we get ⇒ ⇒
Therefore , .
Hence the three parts of 24 are .
4. Find the shortest distance of the point (2, 1, -3) from the plane 2x + y = 2z + 4
Solution:
Let the foot of the perpendicular from the point (2, 1, -3) to the plane be
be .
Shortest distance from (2, 1, -3 ) to the point on the plane is the perpendicular
distance

We have to find the minimum distance equivalently subject to the constraint

.
Let
and (1)
Let =
implies ⇒ (2)

implies ⇒ (3)

implies ⇒
(4)

From (2), (3) and (4) we get

Consider ⇒ ⇒
Consider ⇒ x ⇒
Substituting this in (1) we get + =4 ⇒

⇒ i.e., and

Shortest distance from (2, 1, -3 ) to the point on the plane is given by

5. Find the points on the surface nearest to the origin

Solution:
Let the point on the surface , which is nearest to the origin be .
Distance from this point to the origin is
(1)
But (2)
Using (2) in (1), we get

Now
 r = t=

To find the point at which maximum and minimum occurs we equate

(3)

(4)

Solving (3) and (4) we get

Substituting for in (2) we get
Therefore stationary points are ( 0, 0 , 1) and 0, 0, -1).
At the stationary point (0, 0, 1) the function has a minimum at (0, 0,
1)
At the stationary point (0, 0, -1) the function has a minimum at (0, 0,-
1)
Hence the points on the surface nearest to the origin are ( 0, 0 , 1) and 0, 0, -1).