Math 31A Discussion Session

Week 6 Notes
February 9 and 11, 2016
This week we’ll discuss some (unsurprising) properties of the derivative, and then try to
use some of these properties to solve a real-world problem.

The Mean Value Theorem

First let’s recall one way the derivative reflects the shape of the graph of a function: since
the derivative gives the slope of a tangent line to the curve, we know that when the deriva-
tive is positive, the function is increasing, and when the derivative is negative, the function
is decreasing. When the derivative is zero, it’s hard to really say whether the function is
increasing or decreasing, but we can at least say that the graph is “flat” near the point where
the derivative vanishes.

We can put this to use in the following, relatively unexciting situation: suppose we have
a function f which is continuously differentiable on the closed interval [a, b], and suppose
that f (a) = f (b). If f 0 (c) > 0 for some point c ∈ [a, b), then f is increasing at c; in order
for us to have f (a) = f (b), f must decrease at some point in the interval (a, b), so we have
f 0 (d) < 0 for some d ∈ (c, b). Then the intermediate value theorem says that f 0 vanishes
somewhere in the interval (c, d). Similarly, if the derivative is negative anywhere in (a, b), we
can find a point in the interval (a, b) where the derivative vanishes. So one way or the other,
the derivative vanishes somewhere in the interval (a, b). The following theorem formalizes
and generalizes1 what we’ve demonstrated:
Theorem 1 (Rolle’s). If f is a continuous function on the closed interval [a, b] which is
differentiable on the interval (a, b) and f (a) = f (b), then the derivative f 0 vanishes at some
point in the interval (a, b).
Note. The fact that f is continuous on a closed interval is important. Consider the function
 −1
x − 1, x 6= 0
f (x) = .
0, x=0
This function is differentiable on the open interval (0, 1) and f (0) = f (1), but the derivative
is never zero. Using a closed interval forces f to behave more nicely.
Example. Let f (x) = x2 − 4x + 7. Then f is continuous on the interval [0, 4] and differ-
entiable on (0, 4), and f (0) = f (4). Rolle’s theorem guarantees a point c ∈ (0, 4) so that
f 0 (c) = 0, and we can find this c. We have
f 0 (x) = 2x − 4,
so if f 0 (c) = 0, then 2c − 4 = 0. So f 0 (2) = 0.
1
The generalization is that the theorem only requires the function to be differentiable; our discussion
required the function to be continuously differentiable.

1
 Rolle’s theorem has a nice conclusion, but there are a lot of functions for which it doesn’t
apply — it requires a function to assume the same value at each end of the interval in
question. We can remedy this, though. Suppose f is continuous on [a, b] and differentiable
on (a, b). Then we can define a function g which traces out the secant line passing through
(a, f (a)) and (b, f (b)):
f (b) − f (a)
g(x) := (x − a) + f (a).
b−a
Now consider the function h given by h(x) := f (x) − g(x). Then h is continuous on [a, b],
differentiable on (a, b), and h(a) = h(b) = 0. So Rolle’s theorem says there’s a point c ∈ (a, b)
so that h0 (c) = 0. That is,

f (b) − f (a) f (b) − f (a)

0 = h0 (c) = f 0 (c) − g 0 (c) = f 0 (c) − ⇒ f 0 (c) = .
b−a b−a
The geometric interpretation is this: given a differentiable function f on the interval [a, b],
we can find a tangent line to f between a and b which is parallel to the secant line passing
through (a, f (a)) and (b, f (b)). This is called the Mean Value Theorem.

Theorem 2 (Mean Value). If f is a continuous function on the closed interval [a, b] which
is differentiable on the interval (a, b), then there’s a point c ∈ (a, b) so that

f (b) − f (a)
f 0 (c) = .
b−a

Example. Let f (x) = x3 − 4x. Show that there is precisely one c ∈ (−2, 1) which satisfies
the conclusion of the mean value theorem on [−2, 1].

(Solution) The mean value theorem says that there is some c ∈ (−2, 1) so that

f (1) − f (−2) −3
f 0 (c) = = = −1,
1 − (−2) 3

so we’re looking for a c ∈ (−2, 1) so that

−1 = f 0 (c) = 3c2 − 4.

That is, 3c2 = 3, so c = ±1. We see that c must be -1, since 1 6∈ (−2, 1).

Example. Use the mean value theorem to show that

√ √ y−x
y− x< √
2 x

whenever 0 < x < y. (For what it’s worth, I don’t like this example, but it’s of a type that
I’ve seen asked on midterms before.)

2
 √ √
(Solution) There’s an easy way to do this: observe that ( x− y)2 > 0 whenever 0 < x < y,
so
√ √
0√< x − 2 x y + y
√
2 x y − 2x < y − x
√ √ y−x
y− x< √ .
2 x
But we’re asked to use the mean value theorem, so let
y−x √ √
f (x) = √ − y + x.
2 x
We want to show that f (x) > 0 for all 0 < x < y. Notice that f (y) = 0, and suppose
f (x) ≤ 0 for some 0 < x < y. Then the mean value theorem says there’s a point c ∈ (x, y)
so that
f (y) − f (x) f (y)
f 0 (c) = ≥ = 0.
x−y x−y
But √ √ √
0 2 c(−1) − (y − c)(c−1/2 ) 1 − c − c−1/2 y c −y
f (c) = + √ = + = √ .
2c 2 c 2c 2c 2 c
Since y > 0, f 0 (c) < 0 for all x < c < y, so there’s no x with f (x) ≥ 0.

Extreme Values
Next, we would like to identify the extreme values of continuous functions. The following,
important fact will be needed:
Theorem 3 (Extreme Value). If f is a continuous function on [a, b], then there are values
m and M so that
m ≤ f (x) ≤ M,
for all x ∈ [a, b].
This theorem guarantees the existence of extreme values; our goal now is to find them.
As with the mean value theorem, the fact that our interval is closed is important. The
function f (x) = x−1 is continuous on the interval (0, 1), but does not assume a maximum
value on this interval.

From our discussion at the beginning of the previous section, we see that when f 0 (x) > 0,
f is increasing and when f 0 (x) < 0, f is decreasing; in either case, f (x) is not an extreme
value, because we can move to the left or right and find a larger or greater value. So when
f is differentiable, the only interior points of our interval where a function might assume
an extreme value are those points where the derivative vanishes. The function could also
assume an extreme value at a point where the derivative doesn’t exist; we name the points
where either of these things happens.

3
Definition. A number c in the domain of a function f is called a critical point of f if
either f 0 (c) = 0 or f 0 (c) does not exist.

Of course, just because c is a critical point doesn’t mean that f (c) is an extreme value.
Our graph could have multiple “peaks”, one higher than another, or it could also be the
case that the derivative vanishes without the function taking a local maximum (think of
the graph of f (x) = x3 at x = 0). So the critical values (the value a function takes at a
critical point) become candidates for extreme values. We also must consider the endpoints
of our closed interval, since these could lead to extreme values without having a vanishing
or non-existent derivative.
Theorem 4. Suppose f is continuous on the closed domain [a, b], and let c ∈ [a, b] be such
that f (c) is the minimum or maximum value of f on [a, b]. Then c is either a critical point
of f or one of the endpoints a or b.

Example. Find the maximum value of f (x) = x1/2 (1 − x)3 on the closed interval [0, 1].

(Solution) We have
1
f 0 (x) = x−1/2 (1 − x)3 + 3x1/2 (1 − x)2 (−1),
2
by the product and chain rules. Notice that f 0 (x) is defined on (0, 1], but not defined at 0.
Now if c ∈ (0, 1) is a critical point, then f 0 (c) = 0, so
1 −1/2 1
x (1 − x)3 = 3x1/2 (1 − x)2 ⇒ 1/6(1 − c) = c ⇒ c= .
2 7
Now we must compare the value of f at c to the value of f at the endpoints:
  r  3
1 1 6
f (0) = 0, f = , f (1) = 0.
7 7 7

Whatever the value of f (c), it is positive, and is thus the maximum value of f on [0, 1].

In some cases we can actually determine whether a critical point will give a local maximum
or a local minimum in a somewhat easier way, using the first derivative test for critical
points. For this, let c be a critical point of f . Then if f 0 (x) is positive just to the left of c
and negative just to the right of c, f (c) is a local maximum. On the other hand, if f 0 (x) is
negative just to the left of c and positive just to the right of c, then f (c) is a local minimum.
If f 0 (x) does not change signs at x = c, then f has neither a local maximum nor a local
minimum at x = c.

Example. Consider the function f (x) = 2x+1x2 +1

. Find the critical points of f and the intervals
on which f is increasing or decreasing. Use the first derivative test to determine whether
the critical point(s) give local minima or maxima.

4
(Solution) First we use the quotient rule to find that
(x2 + 1)(2) − (2x + 1)(2x) −2x2 − 2x + 2 −2(x2 + x − 1)
f 0 (x) = = = .
(x2 + 1)2 (x2 + 1)2 (x2 + 1)2
Since (x2 + 1)2 > 0 for all x, f 0 (x) exists for all x. So if c is a critical point of f , then
f 0 (c) = 0. But this implies that
√ √
2 −1 − 5 −1 + 5
c +c−1=0 ⇒ c1 = and c2 = .
2 2
So f has two critical points, one positive, one negative. We can find the intervals on which
f is increasing or decreasing by checking the sign of the derivative in the intervals with
endpoints at the critical points. In particular, −2 < c1 and f 0 (−2) < 0, so f must have
negative derivative at all points less than c1 . That is, f is decreasing on (−∞, c1 ). Similarly,
2 > c2 and f 0 (2) < 0, so f must have negative derivative to the right of c2 . So f is decreasing
on the interval (c2 , ∞). Finally, c1 < 0 < c2 and f 0 (0) > 0, so f is increasing on (c1 , c2 ).
Since f 0 is negative just to the left of c1 and positive just to the right of c1 , f (c1 ) is a local
min. On the other hand, f 0 is positive just to the left of c2 and negative just to the right of
c2 , so f (c2 ) is a local max for f .

Applied Optimization Problems

We’ll conclude by applying our discussion of extreme value problems to two realistic scenar-
ios.
Example. A company that makes sprocket sets for bicycles sells these sprocket sets to a
bicycle manufacturer in bulk for $20 per unit. The total cost of producing x units is given
by
C(x) = 50, 000 + 8x + 0.0003x2 ,
measured in dollars. In a given week, the manufacturer is willing to purchase up to 30,000
sprocket sets at $20 per unit. How many sprocket sets should the production company make
each week in order to maximize their profit?
(Solution) We assume that the production company will not make more than 30,000 sprocket
sets per week, wanting every set to be bought by the manufacturer. Then the revenue
generated by x sprocket sets is
R(x) = 20x,
where 0 ≤ x ≤ 30, 000. So the profit from producing x sprocket sets is
P (x) = R(x) − C(x) = 20x − (50, 000 + 8x + 0.0003x2 ) = 12x − 50, 000 − 0.0003x2 ;
we want to maximize this function on the interval 0 ≤ x ≤ 30, 000. First, we find any critical
points:
P 0 (x) = 12 − 0.0006x ⇒ 0 = 12 − 0.0006c ⇒ c = 20, 000.

5
So the maximum profit will occur when the company produces 0, 20,000, or 30,000 units.
We check these production levels:

Production level (units): 0 20,000 30,000

Profit (dollars): -50,000 70,000 40,000

So the company will maximize its profit by producing 20,000 sprocket sets per week.

Example. A marine energy company plans to place a tidal turbine in the ocean, 5km south
of a straight shore that runs east-west. The turbine must be connected via a power line to a
transformer that is located on the shore, 12km east of the point on the shore that is nearest
to the turbine. The cost of installing this power line is $30, 000 for each kilometer of line
that is in the water and $20, 000 for each kilometer along the shore. Assuming the line will
follow a straight line to the shore (which is not necessarily perpendicular to the shore) and
then follow the shore to the transformer, what is the minimum cost of installing the power
line?

(Solution) Let x represent the distance between the point where the power line hits the
shore and the point on the shore which is closest to the turbine. Then the part of the power
line that is in the water is the hypotenuse of a right triangle
√ with legs of length x and 5km,
so the length of the line which is in the water is given by 25 + x2 . The length of the part
of the line which runs along the shore is given by 12 − x, so the total cost of the line is given
by √
c = (30) 25 + x2 + (20)(12 − x),
measured in thousands of dollars. Notice that c is defined on the closed interval [0, 12]. We
see that
dc 30(2x)
= √ − 20,
dx 2 25 + x2
30x
p
so c has a critical point when √25+x 2 = 20. This is easily solved to give x = 25/1.25 =
√
2 5 ≈ 4.47km. Finally, we notice that
√ √
c(0) = 390, c(2 5) = 50 5 + 240 ≈ 351.803, and c(12) = 390.

So the most cost-effective plan for our power line is to reach the shore approximately 4.47km
east of the point on the shore which is nearest the turbine, and the installation of the power
line will cost approximately $351,803.