Chapter 14

14-6-4

This is a straightforward exercise…..

f = x sin y
δf
f x = sin y
=
δx
δ
f xy
= = f cos y
δy x
δ
f xx =
= f 0
δx x
δf
f y = x cos y
=
δy
δ
f yy = f = − x sin y
δy y

14-6-9

Let’s rearrange the given function.

1
xy + yz + xz =

( x + y) z =
1 − xy

1 − xy
z=
x+ y

Keep in mind that 1 – xy = (x + y) z. We will use this expression to simplify any mess in
the following differentiation.
zx:
− y ( x + y ) − (1 − xy )
zx =
( x + y)2

− y( x + y) − z( x + y)
=
( x + y)2

( x + y )( y + z )
= −
( x + y)2

( y + z)
= −
( x + y)

zy:
− x( x + y ) − (1 − xy )
zy =
( x + y)2

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

( x + z )( x + y )
= −
( x + y)2

( x + z)
= −
( x + y)

zxx:
( y + z)
zx = −
( x + y)

z xx = −
[ zx ( x + y) − (1)( y + z )]
( x + y)2
  ( y + z) 
− ( x + y) ( x + y) − ( y + z)
= − 2

( x + y)

2( y + z )
= −
( x + y)2

zyy:
( x + z)
zy = −
( x + y)

 z y ( x + y ) − (1)( x + z ) 
z yy = − 
( x + y)2

 ( x + z) 
− ( x + y) ( x + y) − ( x + z)
= 
( x + y)2

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

zxy:
( y + z)
zx = −
( x + y)

(1 + z y )( x + y ) − (1)( y + z )
z xy = −
( x + y)2

 ( x + z )  
 1 −  ( x + y) − ( y + z)
 ( x + y) 
= − 2

( x + y)

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

14-7-2

This is a straightforward exercise.

f (x, y) = x2 – y2 + 6x – 10y +2

fx (x, y) = 2x + 6

fxx (x, y) = 2

fy (x, y) = –2y – 10

fyy (x, y) = –2

fxy (x, y) = 0

The discriminant for a point (xo, yo) is

D (xo, yo) = fxx (xo, yo) fyy (xo, yo) – fxy (xo, yo)2

= (2) (–2) – 0

= –4

Since D (xo, yo) < 0, there is neither a minimum nor a maximum at a point (xo, yo).

Courtesy of 3D Surface Plotter

14-7-9

Based on the USPS regulation, we have three variables, in terms of the length, width, and
height:

 + 2 w + 2h =
130

We can reduce it to two variables by expressing the length in terms of width and height:

 = 130 − 2 w − 2h

The volume, V, of the package is

V = wh

= (130 − 2 w − 2h) wh

= 130 wh − 2 w2 h − 2 wh 2

Now, it is time to determine Vw and Vh:

Vw = 130h − 4 wh − 2h 2

Vh = 130 w − 2 w2 − 4 wh

To determine the extremum, we set the two equations above to zero, and solve for w and h.
With some algebra, we get

w = 65/3, h = 65/3, and  =130/3

And the volume of this package is V = (65/3) (65/3) (130/3) ≈ 20343 cubic inches.
 What we get is the extremum. But how do we know that it is a maximum? Well, we can
easily get a minimum, by setting one or two of the three variables to zero! Therefore, what we
have should be a maximum. By the way, we can always take the trouble of calculating the
discriminant at (65/3, 65/3).

Well, a better way to convince anyone is a look at the plot of the function. The maximum
appears to be at (65/3, 65/3). Seeing is believing.

Courtesy of 3D Surface Plotter

14-7-13
Warning: This is a tedious exercise of algebra and trigonometry. Please be patient.

First, since the trough is uniform, we can reduce the problem into a two dimension one. We
can just consider the cross-section of the trough.

Then, it is necessary to label all relevant parts properly, as indicated in the figure above.

Next, let us label and define the many segments and angle in this exercise. We have

2x + y = 2, which give us y = 2 (1 – x)

h = x cos θ

d = x sin θ

The cross-sectional area of the trough is

1 1
A( x, θ ) =hy + hd + hd
2 2

1 1
= ( x cos θ ) [ 2(1 − x) ] + ( x cos θ )( x sin θ ) + ( x cos θ )( x sin θ )
2 2

2 x(1 − x) cos θ + x 2 sin θ cos θ

=

The partial derivatives, Ax and Aθ are

2(1 − 2 x) cos θ + 2 x sin θ cos θ

Ax =

−2 x(1 − x) sin θ + x 2 (cos 2 θ − sin 2 θ )

Aθ =

= 2 x( x − 1) sin θ + x 2 (1 − 2sin 2 θ )

These are the two equations for extremum when we set them to zero. We will then be able to
solve for both x and θ.
 It will be quite a mess. We need to be patient.

The derivative Ax appears to be a simpler one to tackle, and let’s go for it first. When we set
Ax = 0, we have

2(1 − 2 x) cos θ + 2 x sin θ cos θ =

0

cos θ [2(1 − 2 x) + 2 x sin θ ] =

0

2 x sin
= θ 2(2 x − 1)

2x −1
sin θ =
2

We have a relatively simple expression for sin θ, and we will use it to our advantage.

We will substitute this expression for sin θ in our other derivative setting to zero, i.e., Aθ = 0.

2 x( x − 1) sin θ + x 2 (1 − 2sin 2 θ ) =
0

 2 x − 1 2   2x −1  
2

2 x( x − 1)  + x  1 − 2   =0
 x    x  

2( x − 1)(2 x − 1) + x 2 − 2( x − 1) 2 =0

(4 x 2 − 6 x + 2) + x 2 − (2 x 2 − 4 x + 2) =
0

3x 2 + 2 x =
0

x(3 x − 2) =
0

2
x=
3

Now, we can resolve the other variables. Naturally,

y = 2 (1 – x)

= 2/3

and
 2x −1
sin θ =
x

1
=
2

π
θ= or 30o
6

It should be recognized that the bent-up angle should be (90o – 30o) or 60o.

To sum up, the sides and bottom should be 2/3 meter, and the side should be bent up by π/3 or
o
60 .

We can see that there is a maximum by (0.67, 0.52).

Courtesy of 3D Surface Plotter

14-8-2

Is this a classic case of déjà vu? Well, we had tackled Ex-14-7-9 in a “messy” way, and we
are taking care of it in a much more elegant way via the method of Lagrange Multipliers! Here
we go……
 Remember the beloved USPS package? Now, we are going to maximize the volume V with a
given constraint, which is  + 2 w + 2h =
130 . In the Lagrange’s format, we have

V = wh
g (, w, h) = + 2 w + 2h − 130 =0

We have a Lagrange multiplier, λ, such that

∇V =λ ∇g

where

∇V = wh, lh, lw
∇g = 1, 2, 2

Now, we have

wh = λ
h = 2λ
w = 2λ
 + 2 w + 2h =
130

If we multiplier the first equation by  , and the second equation by w, we get

wh = λ 
hw = 2λ w

By equating hw , we have  = 2w. Then, by looking at the earlier second and third equation,
we can conclude that h = w. With minimal work, we can conclude that

 = 2w = 2h

or  = 130/3, w = 65/3, and h = 65/3. Of course, V = 2(65)3/27 ≈ 20343 cubic inches.

Yes, this solution is a lot simpler and elegant that the one we have for Ex-14-7-9. Don’t you
agree?
14-8-12

Déjà vu? The trough again?

Remember the previous Ex-14-7-13 earlier? Let us refamiliarize with the set-up.

However, we should keep in mind that the angle has been switched from inside to outside.
We have ϕ instead of θ as in Ex-14-7-13, or to an earlier exercise Ex-6-2-13 in section 6.2.

Note:

This instructor regrets the possible confusion to students in this class because of the
change in the definition of the “bending” angle. It is suspected that a weak attempt was made
to reconcile the geometric difference between Ex-6-2-13 and this exercise by asking for the
angle to be “bent up” in Ex-14-7-13. In retrospect, it would be much easier to keep the
original definition of the interior bending angle, θ, throughout these three exercises.
Nonetheless, this is just the opinion of this instructor.

Therefore, we have new definition of both h and d:

h = x sin ϕ

d = x cos ϕ
 The cross-sectional area of the trough is

1 1
A( x, y, φ ) =hy + hd + hd
2 2

1 1
( x sin φ ) y + ( x cos φ )( x sin φ ) + ( x cos φ )( x sin φ )
=
2 2

= xy sin φ + x 2 sin φ cos φ

And the constraint is 2x + y = 2, which is translated into the following Lagrange format,

g (x, y, ϕ) = 2x + y – 2 = 0

We should note that the constraint is a function of ϕ even though it is not explicitly expressed.

Now, it is time for the gradients, ∇A and ∇g :

=Ax y sin φ + 2 x sin φ cos φ

Ay = x sin φ
A=
φ xy cos φ + x 2 (1 − 2sin 2 φ )
∇g = 2, 1, 0

With the Lagrange multiplier, λ, we get

y sin φ + 2 x sin φ cos φ =

2λ
x sin φ = λ
xy cos φ + x 2 (1 − 2sin 2 φ ) =
0

Let us work with the third equation and solve for y in terms of x and ϕ. We get

 2sin 2 φ − 1 
y= x
 cos φ 

We will put it together with the second equation back into the first equation to get

 2sin 2 φ − 1  
x   sin φ  + 2 x sin φ cos φ =
2 x sin φ
 cos φ  

2sin 2 φ − 1
+ 2 cos φ =
2
cos φ
 (2sin 2 φ − 1) + 2 cos 2 φ =2 cos φ

2(sin 2 φ + cos 2 φ ) − 1 =2 cos φ

cos φ = 0.5

π
φ= or 60o
3

When we go back to the earlier equation relating x and y, we have

 2sin 2 φ − 1 
y= x
 cos φ 

 2sin 2 (π ) − 1 
= 3 x
 cos( ) π
 3 

=x

Basically, we have recovered the same result as in Ex-14-7-13. Now, we have

x = y = 2/3 meter, and ϕ = π/3 or 60o.

The method of Lagrange multiplier does make life a bit easier, but not by much.

14-8-16

We are given a function f such that

f (x, y, z) = x2 + y2 + z2

and a constraint, which is x + y + z = 9.

Now, we are asked to minimize the function, f, given the constraint,

g (x, y, z) = x + y + z – 9 = 0
 First, we should determine the appropriate gradient.

∇f = 2 x , 2 y , 2 z
∇g = 1, 1, 1

This is a relatively simple problem using Lagrange multiplier. We have the following two
relations:

2x = λ, 2y = λ, 2z = λ

x+y+z=9

From the first relations, we get

2x = 2y = 2z = λ

Together with the given constraint, we have

x=y=z=3