Section 3 .

local
given function
and then determine wether they
Find the critical points of the are

.
saddle points
maxima local minima ,
or
,

7) f(x y) ,
= 3xx +
2xy (x
+

+
y +
y
+ 4

grad f =
(6x 2
2y 1) 8 Critical points
Cy 2x
+ + + + = -
,

2(+) 2y + + 1 = 0
2
2y 2

3
6x + 2y + = 0 6x + +
= 0

1(2x +
2y
+ 1 = 0)t -

2x -

2y
-

1
=
0
=
t +
2y
+ 1 = 0

4X + 1 =
0

= X =

= =
y
=

- =
Thos Y (t, 1) critical point
.
-

,
= is the only

of == =

off-2 , of
1270
(6)(2) 0 =
=>
D(Yo)
-

local minimum
Thus ,
Yo =
(I , I) is a

9) f(x y) , = cos(x* + y') (consider only the three critical points 10. 0 .
1 vol , ,

10 , Fil]

i) Y = (0, 0)

of-2xsinty) d =
-sity) -

xy

= -Cysinty-sixty) dyoy() -
= 0

= -xycos(tyr 0

>
- D(x) = 0 - The test fails
 Consider that -

11COSE11 FOER
.

) cos(0) 1
wos(x : +
y
: =
=

> cos(i +
y2) < 1 = cos(x +
y) F(x , y)ER"
Therefore ,
% .
= 10 ,
0) is local maximum

ii) Yo =
(i ,
Fi)

-sin()-
-0

( =

Gaf
(0) = -

<sin( E) + -

PEcos( E) + = ai

Dia

=- 4
1224)" 0 the test fails
(2) (2)
&
=>
D(X)
=
-

VOEIR
Again ,
consider that -11COSE21

cos(x : +
y )
: =
cos (+) = - 1

>
-

cosl -

+c) = -

1 2 cos(x +
yz) f(x ,
y) ti

Therefore, o
=
( VE) is a
local minimum .
,

iii)o =
(0 Fi)
,

() - Isin(0 P10cs(0 ) failsas

test
= +
) -
+ = - the

cos(X * ) cos(it) 1
that y:
= -

+
Consider
=

cos(xi
=>
+
yi) = -
1 < cos(x +
yz) f(x , y)tR2

Therefore ,
Yo =
10 ii)
, is a local minimum .
11) f(x y)
,
=
ecosy

gradf =
lecosy , -exsiny)
0
excosy -e"siny
=

need
=

We
0
-siny
=

want cosy
=

e + 0
EXER - We

don't have the same X intercept

But
, cosy
and -siny

have critical points .

Thus ,
I doesn't any

points of
19) Find and classify al critical

Ey
10
by
ty Ext
+ +

f(x y) +x
-
-

+
=
,

(x2 y b) 5
gradf 5y +
x =
-
-

=
,

1
X(x 1) x 0 x =

x 0t =

0
= = ,
-
x =

2
3, y
=

2) 0 -
3)(y
=

:
6 0 = (y -
-
=
y
y 5y + =

(0 3) 2) (1 3) , (1 2)
thus ,
the critical points are :
, ,
10 ,
,
,
,

of =
-
x
off
byz
=
Ly-5 ,
: ) Y = (0 3) ,

of (03) = -10 ,
of
h
Dy
(8) = 1 ,
D( % )
= -

10

saddle .
Yo = 10 , 3) is a point

ii) Yo =
(0 2),

102) = 0
, 1
,
D

* =
10, 2) is a local minimum
 iii) o
=
(1 3) ,

out (1 3,
=

10. (3) =
1 D =

>
-

To = (1 3) ,
is a
local maximum

iv) %o =
1 1 2)
,

ort(1 , c) = 10
, (2) = 1
,
Do

>
-

To =
(1 2) , is a
saddle point

21) Find the local maxima and minima for

x
y
3yz)e1
-
-

z = (x2 +

(Exe" -3"
*

gradz =
- (x + 3y :
) exe
** Ge(x 3yye +
=

e" +0 FuERe 2x(1 -

x =

3y) = 0 D

Cy(3 -
y -

3yz) =
00

=1
① is 0 for x = 0 >
-

(y(3 3y2)
-
= 0 +
y
= 0
, y
=

x = 1
2x(1 xi) 0 = x = 0
,

for
=

O
-

0 &
O =

is
y

the critical points 10 01 10 1) (0-1) (1 0) 1-1 0)

= are , ,
, ,
, , , ,

-(ex-yx(e-x-by) =
- Pexy elx-3y)-e +

-O (eby(3-xyz)) =

e by -x - 3y -

+
23xsy-y

OCzd(ex y y(3 3y)) 3) 4xy(3

*
-x e" x xyz) xxy)
-

=
-
-
-
=

terms of and y the result for X 11 and

squared
=

Notice that
X
for
so
and have
Exx Eyy we ,

terms 0 if are 0
.
be the And Exy has Xy so
Exy
= X or
y
y
= 11 is
going to same .
 i) Yo 10, 0)
=

100 Old ,
z
= 220 = be , .
,

= D(o) = 120 0

>
-

Yo = 10, 0) is
local minimum

ii) Yo =
10, 11)

ot (0 , 1) = -20
, (, 1) = 12
,
=> D(Y) = 48

>
-

Yo =
(0 11) are
local maxima .
,

iii) Jo =
(11 0)
,

(1 ,
0 =- 0

, (10) =
4
,to
=> D(Y ) .
= -

16

Yo =
(11 0) ,
are saddle points
.

critical points of
45) Determine the nature of the

f(x y)
,
=

xy
+
+ +

grad f =

(y- * )
*
-
not solution
t = X = 0 is a
0 -

S
-
Y
-
= -
y

- 4
y x)1 8x)
+ -
0
=
=
= x 8x = 0 = = 0 + x =
y
x
- -

=
1 1
 => D(Yo) = 4 -1 = 30

local
>
-

Y .
=
(t 4) ,
is a
minimum

Section 3 .
4

5) Find the extrema off subject to the stated constraints

.

f(x y) ,
=
X
, subject to x+
Cyc = 3

x
g(x y) +
2y2
=
,

If 11 0) (2x xy)
Yg
=
=
,
,
,

(1 0)
- ,
=
1)2X0
, 4 yo)

3g(z)
= 2xx0 = 1 = 3

4x %0
=
0
Yo
= 0
X =
3 Xo = 15

[275 =

1 e X
=
=> Yo
o
=
15 0) ,
,
=
( -
55 , 0)

x" =
3 -

2y" = X = 13 =

2y)
larges value for X is when y = o
Since
y
is
squared ,
the

>
-

xe(- 5) ,
= f(x y)t(
,
+

r, ]

This Y = (v, O) is a local maximum

,

Yo local
=
(5) 0) ,
is a minimum
 of f :"
S ((x y z)/
11) R (x
y z)
x
+y" z2
=

Find the relative

+
extrema >
-
,
, ,
+
e
,
, ,

z42 + xz + y2]

f (x y , , z) = xi + y+ z - Xf = (ax zy 2z) , ,

y (
zy 1)
El z x2
g(x >
(g
-

=
-

y
=
2x
-

- -

, ,
, ,

(2x 2z) x ( =2x 2 y 1) z 2 x

y
+
2
y
+
-

=
, , =
, ,
,

= Ye = 1
,
or % .

x = -1 : 27 =
1 z
= ty = 2 + x +
y -t = x +
y - not valid

=>
0 - 2 0 0 + zo
= 2
Xo :

Y= To =
+ +

= To is a relative extremum .

defined dise namely D S(x g)

13) function flxy) the
=

Consider the X* y
i

+ on
+
xy
= ,

to locate and
lagrange multipliers
the maximum minimum
method of
y2 =13
X" + .
Use the

this to determine the absolute and

points for f on the unit circle .
Use maximum

f D
.
minimum values for on

f(x y) ,
= x "
xy y
+ - Yf = (ex +
y ,
Cy + x)

"
y (2x zy)
g(x y) ,
= x = Y
g
=
,

2xx]3x
2x +
y
=

-
3(x 2x(x y) =
+
+
3y = 2xx +
2xy = +
y) =
+
X

2y + x =
2xy
xi +
y = 1

- 2x + = 3x y
= x = x =

y
=

m f( ) =
y

Yg =
0 for Y .
= 10 0)
, so it should be included as possible .
extrema
 · lau =

20
>
-

D(x) = 4 -
1 =
30

local minimum
Since YotS ,
then 10, 6) is a

f(0 0) ,
= 0

E1 and the absolute minimum is .

0
So the absolute maximum is

17) Find the maximum and minimum values attained by flxyiz) =

xyz on the

unit ball xi +
y + z

Yf =
(yz , xz ,
xy)
#
g
=
(2x Cy 2z)
, ,

>
-
yz 2xxz(x 2x

&
=
y) + =

z
+ x)
y(z
x) 2x
=
+

xz =
2xY
* y) 2x
y) =
z +

2x7
xy
=

X z 1 1
+
y + =
= 3(2x)" = 18 x = =
25

= .
= =
) iii)

f(i) =
(i)" = >
- maximum

f(x) =
1 j)" -1 =

,
e minim is

19) a) Find three numbers whose is 27 and whose som and whose som is

minimal

zl
 => 1 Jyz x2 27x2z

S
=
=> 1 =
xyz =
x2(xyz)z =

1 = XXz
1 =
X2 xyiz =

27x2y
e X =
y = z =

2
1 =
xxy 2
1 X 27xx
2
=
x
yz
=

27
xyz =

=1 =
=>
xyz

= 1 1
x =
=

%
27 9

81
=> Xo zo = 3
Yo
= = =

27

E Xo zo = 3
yo
= =

b) Find numbers whose som is 27 and whose and product is maximal

three

f(x
zy
S
, y z) ,
=
xyz

z)
g(x y x +
= +
, ,
y

& &3
2x
yz
=

x z
=> y
= =

xz
= 2x

=
2x
xy

27 27 - 3x 27 Xo =
yo =
zo =
9
+ + z
=
x
y
x =
+ x + x + =

critical of constraint 2x2 y2 1

.
33)a) points subject the + =

to
Find the x +
ye ,

Yf =
( 2y),

Tg
=
(4x ,
2
y)

- 1 = 4xX

cy =
1xy + x = 1 = Xo =
t
2x + y =

-y = 1 -

! = yo
= = v
x =
i
==
(4 = ) =

* + 1
yo
= 0 = (x = 1 = xo = =
/ x =
=

W=
 = =

( =
H 0) ,

Hessian to classify the critical points.

b) Use the bordered

x((x yz) (xx (1 x)y

f xg
= x +
y + = X + -

h(x
-
-

y)
-

=
,

on

OX
= 1 - 4xx =
= -
4x

82h
= O
0x0 %

Oh =
c(1-ye
O =

0 -

4x -

2y

>
-
IH) = -

4X -

4X O

= 4x( -

8x(1 -

x)) -

2y( 8yx)
-

2y
-

O 2(1 x)
-

32x4(1 x) 16yX
=

-
-
+

Yo =
/+1 /E has x =

(t =)
local maxima
14)0
are

(H) =
()
=

- =

= ( 0) ,
has i
=

-It =
-

x (d) 20- =
(10) is a
local minimum

To =

(= 10) was n
=

=> IE = -

32 (t)) 10 %
=

(-10) is a
local minum .
 (cost sint sin (1)) that is farthest from the
37) Find the point on
the curve , , origin .

distance of point (x % 2)
, ,
from the
the
f = x = y + z2 is

onigin
sin't + Sin = 1 + sin't
cos2t +
& f(t) =

> = in() cos( sint

>
- t = ut

sin() The period is 4th

considering
>
-
z =

distance
>
-
We have four cases for the largest

t =
0 i zi , 37
,
,

f(0) = 1

2
f(m) = 1 + 1 =

f (2m) =
1

1+ 1 2
f(z)
=
=

So the maximum distance will be at

* =
(1 0
,
-1) or
Yo = ( -
1
,
0
,
-1)
,
Bonus

. 3
3 31) Write 120 of three numbers so that the som of the products
.
number as a som

take two at a
time is a maximum .

x +
y
+ z =
1207z 120 X
y
= -
-

f(x ,
y z)
,
=

xy
+ xz +
yz

>
-

f(x ,
y) =
x(y + 120x -
x2
Yy
- +
120y
-

xy
-

ye

=
120X +
120y -
x2 -

xy
-

ya
>
-
Yf =
(120 -

(x 120 x) =
y 2y
-

- -

120 (x
120 2y
3
0
10
=

y x
-
-
=

2x
=

y
-

=
-

0
180 x =

2y
- =

> = X
y
-

120 40
3x = 0 X
y
=
- = =

z = 120 -

40 -
40 = 40

=> Yo E 40
Yo
= = =

3 . 3 .

47) Let u be a harmonic function on D Show that if u achieves its

.

maximum value in DLOD ,

it also achieves it on
OD .
This is some times

called the "weak maximum

principle" for harmonic functions .

Let's consider /U + ene") En o

,

Let SE ?? be a
sequence s . t
. limE =
0 and 507E ... Em

Given that
*
30 Exin then Ene LO FM x
En 2
,

,
,

(Ene ) (En 0)
*
* = ex ,
0
,

# (Ene ) *
= Ene (0
 (U Ene") Ene") 0
"

From that + =

this ,
we
get

retent is strictly sub harmonic

know that
From the previous problem we

OD
which that it has maximum on
implies a

Ene.
Consider the notation Un = U +

of Ur and (x y ) EDLOD the

Let 1Xn , yu)EOD be a maximum , ,
.
be

maximum for U .
Since Un =

Util ,
then Un)U FX

Thus U/X0
, , yo) < Un(X0 yo) ,
< Un (xm yu)
,

Considering the hint , we know 7 (xq Yg) =

,
s .
t .

every neighborhood of

a contains at least 1 member of the

sequence [ph]

close to
This , J /xq Yg) =

,
on
OD Sit g is arbitrarily

that Un- > U

(Xu yu) for k As ke 8, we
get
any
.

So
UIE) =
12() > ((x0 , yo)

Thus its entries a

=
aje.
3 4
. .

31) Let A be a nonzero symmetric 3x3 matrix .

,

Consider f(x) = +(A * ) .

a) What is Yf( * ) ?

f (5) =
ta,anAis X*+ AnAznAnsig &Ais = A23 Abs z

If
>
-
=
(a ,
Giz9isX ,
dir Azz92s
Y ,
9002 as . z) = A *
 b) Consider the restriction of f to the unit sphere
.
By theorem 7

and minimum S Show

maximum on

thatI must have

.

a
we know

t AY X *
be itS and XO s
=

that must
.

there
.

an

** 0
(2x 2z) 2 -where
Xg
=

x
y+ z <y
g(x
=

,
y ,
z) = + , ,

multipliers we get
Using lagrange
,

the

If =
xg + Ax =
( = xx

Since 3 then JES .

and
st
we have maximum minimum on ,

58.
A = 11 % with 10,

c) the maxima and minima for f on B =

[(x y z)(x +
y =-+
13 ?
What
, ,

are

+ *. also have : .
8
A
=

have
=
But we

Again , we

3 .
4 .

39) Carry out the analysis of example 10 for the production function

AKL"" constants and

Q (K 2),
=
,
where A and a are
positive OLX1 .

function and is sometimes used

This called a lobb-Douglas production as a

simple model for the national economy

.

we consider that glk , 2) =

pl + qk =

B with g10

KAK "L"" 11-a/AK * )

*
=> YQ =

, Yg
=
(9 P) ,
,

XAk
*
"L" =
Ap
*
11-x)Ak"- = &q

qk
-

B = 0
pl +
 A***
E

11-a)AU
=

>& =
=
-

q
P

+PAL -
PL

1 x
- p(i -y) +
)
-

= B = L =
B
P
1 -
x

>
-
pB + qk = B

= -+ xB
-
qk

=
- k