Maxima and Minima

Method of finding Maxima and Minima

Method of finding Maximum and Minimum values.

It is also called Stationary or extreme or turning values of a function.

f f
1) solve  f x  0;  f y  0 ( simul tan eously for x and y )
x y
2 f 2 f 2 f
2) 2 . 2  ; for values of x, y obtained in (1)
x y xy
3) when the condition is satisfied ;
2 f 2 f
f  x, y  is max imum , if  0 or 0
x 2 y 2
2 f 2 f
and f  x, y  is min imum if  0 or 0
x 2 y 2

At x  a and y  b; Let us Denote

2 f 2 f 2 f
r  2  f xx ; s   f xy ; t  2  f yy
x xy y

For MAXIMA And MINIMA we must have:

1) if rt  s 2  0; and r  0; then f  x, y  is max imum
2) if rt  s 2  0, and r  0; then f  x, y  is min imum
3) if r  0 (or t  0); then f  x, y  is neither max imum or min imum

Ex1
Discuss the max ima and min ima of x 3  y 3  3x  12 y  40
sol : f  x, y   x 3  y 3  3x  12 y  40
I ) f x  3x 2  3, f y  3 y 2  12;
f xx  6 x, f xy  0, f yy  6 y;

II ) Solve f x  0, f y  0
 3 x 2  3  0;  3  x 2  1  0;  x  1, 1
and 3 y 2  12  0;  3  y 2  4   0;  y  2, 2
 the sationary values are  1, 2  ;  1, 2  ;  1, 2  ;  1, 2 

III ) a ) for x  1, y  2
r  f xx  6(1)  6; s  f xy  0; t  f yy  6  2   12
 rt  s 2  72  0  72  0; and r  6  0
  1, 2  is a min ima
 the min imum value  1  8  3  24  40  22

b) for x  1; y  2;
r  f xx  6; s  f yy  0; t  f yy  6  2   12
 rt  s 2  72  0
 we reject this pair

c ) for x  1; y  2
r  f xx  6; s  0; t  12
 rt  s 2  72  0
 we reject the pair

d ) for x  1; y  2
r  6; s  0; t  12
rt  s 2  72  0  72  0; and r  6  0;
  1, 2  is a max ima
The max imum value   1  8  3  24  40  58

Ex 2
find the extreme values of xy  3  x  y 
sol : f  x, y   3 xy  x 2 y  xy 2
I ) f x  3 y  2 xy  y 2 , f y  3x  x 2  2 xy;
f xx  2 y, f xy  3  2 x  2 y , f yy  2 x;
II ) Solve f x  0, f y  0
simultaneously
 y  3  2 x  y   0 and x  3  x  2 y   0;
 y  0 or y  3  2 x       1
When y=0;

Then
x  3  x  2 y   0; gives x  0 or x  3
The points are
 0, 0  and  3, 0 
When y = 3 - 2x;

Then
x  3  x  2 y   0; gives
x 3  x  2  3  2 x    0
 x  3  3 x   0
 x  0 or x  1
when x  0; y  3  2(0)  3 from  1
when x  1; y  3  2  1  1
The points are
 0,3 and  1,1
Hence the values are
 0.0  ;  3, 0  ;  0,3 ;  1,1
III)
a) When x=0; y=0
r  f xx  0; s  f xy  3; t  f yy  0
rt  s 2  9  ve 
Reject the pair

b) When x=3; y=0

r  f xx  0; s  f xy  3; t  f yy  6
rt  s 2  9  ve 
Reject the pair

c) When x=0; y=3

r  f xx  6; s  f xy  3; t  f yy  0
rt  s 2  9  ve 
Reject the pair

d) When x=1; y=1

r  f xx  2; s  f xy  1; t  f yy  2
rt  s 2  4  1  3  ve 
 f  x, y  has a stationary value at x  1; y  1
 r  f xx  2  0
f has a max imum at x  1; y  1
Maximum value   1  1  3  1  1   1  3  2   1

Ex 3
find the stationary values of sin x.sin y.sin  x  y 
sol : f  x, y   sin x.sin y.sin  x  y 
I ) f x  sin y cos x.sin  x  y   sin x.cos  x  y  
 sin y.sin  2 x  y 
similarly f y  sin x.sin  x  2 y 
f xx  2sin y.cos  2 x  y 
f xy  cos y.sin  2 x  y   sin y.cos  2 x  y   sin  2 x  2 y 
f yy  2sin x.cos  x  2 y 

II ) solve f x  0; f y  0;
 sin y sin  2 x  y   0; and sin x sin  x  2 y   0

 x  0; y  0 or x  2 y   ; 2 x  y   ;
 2 x  4 y  2  multiply by 2  ;

 3y    subtracting  ; hence y 
3
 
 x  2.  ; x 
3 3
III ) a ) x  0, y  0;
rt  s 2  0  0  0; we reject the pair
  3
b) x  ; y  ; r  2. .  1   3
3 3 2
 3 3 9
s ; t   3; rt  s 2  3    0
2 4 4
and r   3  0
 
x ;y is a max ima;
3 3
  2 3 3
max imum value  sin sin sin 
3 3 3 8
Ex4
Divide k ( 0) a real number int o three parts such that their product is max
Also find the max value

sol : let x, y, z be three numbers such that x  y  z  k

let u bethe product. u  xyz and we want u to be max

z  k  x  y, hence u  xy (k  x  y )
f  x, y   xy  k  x  y 

I ) f x  y 1.  k  x  y   x.  1   y  k  2 x  y 
f y  x 1.  k  x  y   y  1   x  k  x  2 y 

f xx  y  2   2 y f xy  1.  k  2 x  y   y  1  k  2 x  2 y
f yy  x  2   2 x
II ) solve f x  0, f y  0
 y  0 or 2 x  y  k and x  0 or x  2 y  k
x  0, y  0 is not acceptable.  product will be zero 

4 x  2 y  2k and x  2 y  k  Multiply by 2 
3 x  k ; ( subtracting )
x  k / 3; y  k / 3
2k
III ) r  2.k / 3   0; k  0
3
4k 2 k 2 11k 2
rt  s 2    0
3 9 9
f is max imum
k k k
z  (k  x  y )  (k   ) 
3 3 3
ERRORS and APPROXIMATIONS
Let f(x,y) be a continuous function of x and y.
Let  x and  y be the increments in twoindependent var iables x and y .
The new value of f  x , y  will be f  x   x , y   y 
 f  f  x   x, y   y   f  x, y 

Expanding By Taylors Theorem and ignoring higher powers, we get

f f
 f  x  y
x y

If we denote f by z , we get
z z
z  x  y
x y
THIS RESULT IS EXTREMELY USEFUL IN CALCULATING ERROR IN Z DUE TO SMALL
ERRORS IN X and Y.
x
if  x is the error in x, then is the relative error
x
100  x
and is the percentage error in x.
x

REPLACE INCREMENTS BY DIFFERENTIALS, WE GET

z z
dz  dx  dy
x y
This Is the Theorem of Total differentials

Ex 1
if ax 2  by 2  cz 2  1 and lx  my  nz  k ; prove
dx dy dz
 
bny  cmz clz  anx amx  bly
dx
also find
dz
sol : let f  ax 2  by 2  cz 2  1  0; and   lx  my  nz  k  0

f f f
 df  dx  dy  dz  0
x y z
 axdx  bydy  czdz  0         1
   
 d  dx  dy  dz  0
x y z

ldx  mdy  ndz  0            2 

From (1) and (2) using Cramer’s rule we get

dx dy dz
 
by cz ax cz ax by
m n l n l m

ax+by+cz=0

px+qy+rz=0

dx dy dz
 
bny  cmz clz  anx amx  bly

dx bny  cmz
 
dz amx  bly

Ex 2

In calculating the volume of a right circular cone errors of 2% and 1% are found in measuring
height and base radius resp. find the % error in calculating volume.

Sol: let r and h be the base radius and height of the cone, then the volume is given by

1
v   r 2h
3

1 
 log v  log     2 log r  log h
3 

1 1 1
 v  2 r  h
v r h

v r h
100  2 100  100
v r h

v
100  2  1  2  4
v

 percentage error in calculating volume  4%

Ex3
Find the percentage error in calculating the area of a rectangle when an error of 3% is made
in measuring each of it sides.

Sol: Let l and b be the length and breadth of a rectangle, then the area of the rectangle is
given by A = l x b.

log A  log l  log b

1 1 1
 A  l  b
A l b

A l b
100  100  100
A l b

A
100  3  3  6
A

 percentage error in calculating Area  6%

Ex 4
Find the possible error in percentage in measuring the parallel resistance r of two
r1 r2
resistances and which undergo a positive error of 2% each in their measurements.

Sol
1 1 1
we have  
r r1 r2
Taking differentials
1 1 1
 r   r1   r2
r2 r12 r2 2

1 100 r 1 100 r1 1 100 r2

 .  .  .
r r r1 r1 r2 r2
1 1 1 1
 .2  .2  2   
r1 r2  r1 r2 
 1 1 1 1
 2.   sin ce  
r r r1 r2
100 r
 2
r
 the error in r is 2%.
Lagrange’s Method of Undetermined Multipliers

To find Extreme values of u=f(x,y) where φ(x,y)=0

Form a new function called Lagrange’s function

F  x, y   f  x, y    ( x, y )

For Stationary Values , we have dF  0

f f    
 dx  dy    dx  dy  0
x y  x y 
 f    f  
     dx      dy  0
 x x   y y 
f  f 
   0;  0
x x y y

Solve the equations.

Ex1
find the stationary values of u  x 2  y 2; subject to
the condition x  y  2; by Lagrange ' s method

sol : u  f  x 2  y 2      1
  x  y  2  0       2
consider Lagranges ' s function

F  x, y   f  x, y     x, y 
 x2  y 2    x  y  2

dF   2 xdx  2 ydy     dx  dy 

dF   2 x    dx   2 y    dy
For stationary values dF  0 ;

dF  0; gives  2 x     0 and  2 y     0       3 
 
from  3 x  ; y       4
2 2

  
from  2  ;   2  0;    2
2 2
 2
from  4  ; we get x    1; y  1
2 2

Hence the point of stationary value is (1,1)

hence f  x, y   x 2  y 2 is maximum at x  1 , y  1
max imum value  12  12  2

Ex2
The sum of the three positive numbers is 1. Determine the maximum value of their product.
Sol
We have to maximize u=xyz
Condition is x+y+z=1.
  x  y  z 1
F  u  
dF  du   d
du  yzdx  zxdy  xydz

d  dx  dy  dz  0

dF  0; gives du   d  0
yzdx  zxdy  xydz    dx  dy  dz   0

 yz    dx   zx    dy   xy    dz  0
yz    0; zx    0; yx    0

xyz   x  0; xyz   y  0; xyz   z  0

 u   x  0; u   y  0; u   z  0.

 3u    x  y  z   0;
3u    0;   3u; (since x  y  z  1)

 u  3ux  0;  1  3x  0;  x  1 / 3
Similarly y=1/3; z=1/3; max value = 1/27

Ex3
If x, y, z are the angles of a triangle , show That 
u  cos x cos y cos z is max imum, when triangle is equilateral

sol : u  f  cos x cos y cos z

condition x  y  z  1800  
   x  y  z   0
consider Lagranges function

F  x, y, z   f  x, y, z     x, y , z 

  cos x cos y cos z     x  y  z   

dF
   sin x cos y cos z  dx    cos x sin y cos z  dy    cos x cos y sin z  dz
   dx  dy  dz 
   sin x cos y cos z    dx    cos x sin y cos z    dy 
  cos x cos y sin z    dz
For stationary values dF  0 ;

dF  0; gives
 sin x cos y cos z    0;  cos x sin y cos z    0;
 cos x cos y sin z    0          3 

From (3); we get

sin x cos y cos z  cos x sin y cos z  cos x cos y sin z  

dividing by cos x cos y cos z; we get

tan x  tan y  tan z; hence x  y  z

Now x  y  z  1800
 hence x  y  z  600