Oct 2 Lecture

Absolute Max & Min

ExtremeValue Thm Let f(x Y)
, be continuous on a
closed bounded domain Df in
and R2 then there exists points
(X, Y1) & (X2 Y2) in Df S t f(x1 Y1) & f(X2 yz)
,
.
.
, ,
are global
max & min values .

# in 1-D on (-0, 8) no global max0

&
min is f(o) =

o
on [-1 , 2) global max is f(2) = &
global min is f(o) o
!
=

-
Technique
global max/min values of f(X y) over closed bod o
find the
To , a

Find the values of the criticals contained

/Bidentify theeroblem)
1 . in D

.
2 Find the max & min values of f along the
boundary of D (Boundary &end points
curve

.
3 Compare. Biggest value =
global max value
smallest value =
global min value
Example
Find the & min values of f(x i y) 2x2 &x + 3y"+ 2
global max
=

on the circular disk of radius / centered at (1 0) ,

Draw D
equation of circle : (x-12 + y =
c

=> y =
=
F
(x)z
-
fx 1
0
= 4x -
4 =
= X =

fy
=

by =
0 = y = 0

Critpoint at (1 , 0)

Boundary f 2x2 4x
: = -
+ 3y2 + 2

4
:

y = N
(x+ )

f(x, f(x) =
2x 2x- + 3)1 -
(x -

1)2) + 2

= 2x2 -
+x + 5 -
3(x -

1)
fx =
+x -
4 -
6(x 1) -
=
0

↑x -

4 -

6x + 6 =
0

=
3 Bandurupt
-
2x + 2 =
0 = X

Y =D x =
y
=
v( 1 -
=

(2 :

Y =
M x
))
get some X value X =
1 as in G
Y = -

XF =
-1 ,
so have demdaypt (1-1)
Endpoints (where boundary curves meet
P
(0 0)
,
& (20)
(0 0) a ·
(2, 0)
,
(9, 0

C
So
,
candidate pt (x y) , +(x , Y)

crit (1 0) , f( 0) . =
0
pt
(1 ,
1) f( 1) ,
=
3

Boundar, , , f( 1) ,
-
= 3

(0 0 , f(0 , 0) =
2
Endpoints
(2 0)
, f(2 0) ,
=
2

Global max value is 3

Global min value is O
Constrained Optimization (Example
Find the biggest the product of X ,
y ,
47 cabe
if 2x +
3y + z = 10 & X30
, y30, z30 .

P =
xyz -z = 10 -
2x -

by
=> P =
Xy(10 -

2x -

3y)
=
loxy -

2xy -

3xy

Px =

10y -

&xy -

3yz =
0

Py 10x 2x2 6xy 0

= - =
-

Factoring ,
PX =
y(10 -

&x-sy) =
0

=> X O or 10 -1x- 3y

5 (10
=
0

=>
y =
x)
*
Since y 0
-

6x( )
*
Plugin to Py Py
*

,
=
10x-2x - =

=> lox-2x2 -
20x + 8x2 = 0

=> -

10x + 6x2 = 0

= ) X(6x 10) - =
0

= Xo or X =
1
%6

Whe 1 %6 1
% &z 103
x =
, y = =

so p =
(i)()() =
1
Lagrange Multipliers
Goal :
Optimize the function f(X Y) , subject to the
constraint g(X y) K. ,
=

Idea :
Suppose f has on extreme value at point P(Xo, Yo, zo)
the surface S & let be a curve
on
given by

F(t) <X(t) Y(t)

=
, ,
Est)) that lies on s & passes through P
.

If F(to) =
<X0 , Yo / zo7 then E(f(CH)) + -t
=
0

By chain rule => Yf(r(to)) ·

r(to) =
0
.e
i. Of (40 Yo , 70) ·to)
,
=

So I
Of F1 at t to = .

Recall that discussing taught places

g(X ,
y, z) =
K satisfies Vg1r/
So

and
Dg(X0 Yo
.
zo)
Of (X0 Yo , to)
.
,
+ GSto)
+ E(to) OfBel at extrema

sent vectors
paralle
-
 Method

1) Solve the system of equations

/Of
= Lagrange
multiplier
(g =
k

2) Plugsolutions of 1) into f to identify max/mins if

they exist &1975 on the surface g = K

Example Find the extreme values of f(x Y)

,
=
X2+ 242
on the circle X + yEl .

Of =
< 2x &y) ,
, Xg =
22x , 247

Solve
a
②
(x2 +
y2 = 1 ③

From D , X= 0 or x =
/

fX ,then saysX =
1)
So points
,
are (0 , +1) , (11 0) ,

!
so max is 2

min is