You are on page 1of 6

LECTURE 16: OPTIMIZATION

1 THE OPTIMIZATION TECHNIQUE

Derivatives can be used to find maxima and minima. If we wish to optimize a system, this usually requires
maximizing some output or minimizing some cost. Hence, derivatives are useful for solving ‘real world’
problems. The process of finding the ‘best setup’ for a system is known as optimization.

Technique:

- Convert the problem to a function (usually of two or more variables)

- Find constrains.

- Use constraints to reduce the problem to a single variable.

- Find absolute extrema.

- Interpret solution.

2 EXAMPLES OF OPTIMIZATION

2.1 An Open Box

A square of paper with sides or 12 cm is cut to produce an open box. What is the largest volume the box
can have?

Figure 1: An open box from a square of paper.

Function to be maximized - Volume:


V = x(12 − 2x)2 = 4x3 − 48x2 + 144x,
with x ∈ [0, 6].

Already a function of one variable. No constraints needed.

Find critical numbers


V 0 = 12x2 − 96x + 144 = 12(x2 − 8x + 12) = 12(x − 2)(x − 6).
Critical numbers when V 0 = 0, hence, at x = 2 and x = 6.

Apply closed interval method to find absolute maximum.

1
Make Table:

x 0 ··· 2 ··· 6

V 0 (x) N/A + 0 − 0

V (x) 0 % 128 & 0

Hence, the maximum volume is when x = 2 i.e. when one cuts squares of size 2 cm × 2 cm out of the
corners of the sheet of paper. In this case the volume of the box created is 128 cm3 .

2.2 A Field by a River

A farmer has 1200 m of fencing and wants to fence off a rectangular field that borders a straight river. The
farmer needs no fencing along the river. What is the largest area that can be fenced off?

Figure 2: A field by a river.

Function to be maximized - Area:


A = xy,

Constraint - Limited fencing:


y + 2x = 1200.
Reduce to function of one variable by eliminating y
A(x) = xy = x(1200 − 2x) = 1200x − 2x2 ,
with x ∈ [0, 600].

Find critical numbers


A0 (x) = 1200 − 4x = 4(300 − x).
Critical numbers when A0 (x) = 0, hence, at x = 300.

Apply closed interval method to find absolute maximum.

Make Table:

x 0 ··· 300 ··· 600

A0 (x) N/A + 0 − N/A

A(x) 0 % 180, 000 & 0

Hence, the maximum area is when x = 300 i.e. when the field is 300 m deep and y = 1200 − 2 × 300 =
600 m wide. In this case the area of the field is A = 300 × 600 = 180, 000 m2 .

2
2.3 A Cylindrical Can

Find the dimensions of a cylindrical can such that, for a given volume, the surface area is minimized.

Figure 3: A cylindrical can.

Function to be minimized - Surface Area:

A(r, h) = 2πr2 + 2πrh,

Constraint - Constant Volume:


V = πr2 h.
Reduce to function of one variable by eliminating h
V V
A(r) = 2πr2 + 2πrh = 2πr2 + 2πr = 2πr2 + 2 ,
πr2 r
with r ∈ (0, ∞).

Find critical numbers


V
A0 (r) = 4πr − 2 .
r2
Critical numbers when A0 (x) = 0, hence
V
4πr − 2 = 0,
r2
4πr3 − 2V = 0,
r
3 V
r= .

As the allowed values are on an open interval (0 is not in the domain of the function A(r) and ∞ is not
a number) we cannot use the closed interval method. The endpoints of the interval do not need to be
considered, hence, we only need to worry about the local extrema. Both the first and second derivative test
can be used to analyse the extrema. For this example we shall use the second derivative test.
V
A00 (r) = 4π + 4 .
r3
p
3
At the extrema, when r = V /2π, we have

V
A00 (r) = 4π + 4 V
= 12π.

p
As A0 (r) = 0 and A00 (r) > 0 at r = 3
V /2π, this point is a minimum.

Hence, the surface area is minimum when r


3 V
r= ,

3
and r
V V 3 4V
h= 2 = 2/3 = ,
πr 
V π
π

and the aspect ratio of the can is r
3 V

r 1 1
= r 2π = √
3
= .
h 3 4V 8 2
π

2.4 Closest Distance Between Two Points I

Find the point on the parabola y 2 = 2x that is closest to the point (1, 4).

Figure 4: Schematic diagram of the point on the parabola y 2 = 2x that is closest to the point (1, 4).

Function to be minimized - distance to point (1, 4):


p p
d = (x − x0 )2 + (y − y0 )2 = (x − 1)2 + (y − 4)2 .

Constraint - point must lie on the parabola y 2 = 2x:


y2
x= .
2
Reduce to function of one variable by eliminating x
s 2
p y2
d= (x − 1)2 + (y − 4)2 = − 1 + (y − 4)2 .
2
with y ∈ (−∞, ∞).
√ √
Note that, as x is always increasing on its domain, the minimum of x is also the minimum of x. Thus,
we can minimize d2 and still obtain the correct answer.
 2 2
2 y
d = − 1 + (y − 4)2 ,
2

4
Hence, critical numbers are
y2
 
(d2 )0 = 2 − 1 · y + 2(y − 4) = y 3 − 2y + 2y − 8 = y 3 − 8.
2
Critical numbers when (d2 )0 = 0, hence, at y = 2.

As the allowed values of y are on an open interval we cannot use the closed interval method. The endpoints
of the interval do not need to be considered, hence, we only need to worry about the local extrema. Both
the first and second derivative test can be used to analyse the extrema. For this example we shall use the
first derivative test

Make Table:

y ··· 2 ···

(d2 )0 − 0 +

(d2 ) & 2 %

The closest point on the parabola y 2 = 2x to the point (1, 4) occurs when y = 2 and, hence, when
x = y 2 /2 = 2. Thus the closest point is (2, 2).

2.5 Closest Distance Between Two Points II

Find the point on the parabola y = x2 + 1 that is closest to the origin.

Figure 5: Schematic diagram of the point on the parabola y = x2 + 1 that is closest to the origin.

Function to be minimized - distance to point (0, 0):


p p
d = (x − x0 )2 + (y − y0 )2 = x2 + y 2 .

Constraint - point must lie on the parabola y = x2 + 1:


y2
x= .
2

5
Reduce to function of one variable by eliminating x
p p
d = x2 + y 2 = y 2 + y − 1.

with y ∈ [1, ∞).

As d is always increasing on its domain, we can minimize (d2 )0 .

d2 = y 2 + y − 1,

Hence, critical numbers are


(d2 )0 = 2y + 1.
Critical numbers when (d2 )0 = 0, hence, at y = −1/2. However, this is not in the Domain of d. Thus, there
are no local minima in the interval.

Apply closed interval method to find absolute minimum.

Make Table:

y 1 ···

(d2 )0 N/A +

(d2 ) 1 %

The closest point on the parabola y = x2 + 1 to the origin occurs when y = 1 and, hence, when x = 0.
Thus the closest point is (0, 1), the minimum of the parabola.

Challenge questions are beyond the level of the course and are for students who wish to try more advanced
problems. They are not assessed and do not contribute the final grade for the course.

Q). For a given perimeter, P , show that the shape that encloses the largest area is a circle.

You might also like