You are on page 1of 8

# Maxima and Minima Lecture

Carmen A. Bruni
1 Introduction
Example 1.1. A Texas based company called Hamilton’s Wares sells baseball bats at a ﬁxed
price c. A ﬁeld researcher has calculated that the proﬁt the company makes selling the bats at
the price c is
P(c) =
−1
2000
c
4
+
1
5
c
3

51
2
c
2
+ 1150c
At what price should the company sell their bats to make the most money?
Intuitively, what would we have to do to solve this problem? We wish to know at what
point c is this function P(c) is maximized. We do not have many tools at the moment to solve
this problem so let’s try to graph the function and guess at where the value should be.
From the graph, we can estimate that we should sell bats at approximately 170 dollars per
bat.
However, this is only an estimate. What if wanted to know the exact value? What if we
didn’t have the aid of a computer? Could we still solve these problems? The answer to this
question is yes and this will be the topic of today’s lecture.
1
2 Absolute/Global Maxima
First let’s introduce some terminology.
Deﬁnition 2.1. Let f be a function deﬁned on an interval I containing c. We say that f has
an absolute maximum (or a global maximum) value on I at c if f(x) ≤ f(c) for all x contained
in I. Similarly, we say that f has an absolute minimum (or a global minimum) value on I at
c if f(x) ≥ f(c) for all x contained in I. These points together are known as absolute/global
extrema.
Let’s compute some examples of these values.
Example 2.2. f(x) = x
2
+ 1 for x ∈ (−∞, ∞) (remember this notation means for x living in
the interval from negative inﬁnity to inﬁnity. This can also be written as x ∈ R or in words as
for all real x).
This function has an absolute minimum of 1 at the point x = 0 but no absolute maximum on
the interval.
Example 2.3. f(x) = x
2
+ 1 for x ∈ [−2, 2] (remember closed brackets means we include the
endpoints in our interval).
2
This function has an absolute minimum of 1 at the point x = 0 and a absolute maximum of
f(±2) = (±2)
2
+ 1 = 5 at the points x = 2 and x = −2.
Example 2.4. f(x) = x
2
+1 for x ∈ (0, 2] (remember open brackets means we omit the endpoint
in our interval).
This function has no absolute minimum and a absolute maximum of 5 at the point x = 2.
Example 2.5. f(x) = x
3
for x ∈ (−∞, ∞)
This function has no absolute minimum and no absolute maximum.
3
3 Extreme Value Theorem
So based on these examples, when does a function have an absolute maximum and mini-
mum? These examples seem to suggest that if we have a closed interval then we’re in business.
The following example shows that this is not suﬃcient in all cases.
Example 3.1. Consider the function
f(x) =

x if 0 < x < 1
1.5 if x = 0, 1, 2
−x + 4 if 1 < x < 2
Graphing yields
From the graph, its clear that this function has no absolute minimum or absolute maximum
but f(x) is deﬁned on all of [0, 2]. The problem with this example is that the function is not
continuous.
Theorem 3.2. (Extreme Value Theorem [EVT]) Let f(x) be a continuous function
deﬁned on a closed interval. Then f(x) has an absolute maximum and an absolute minimum
on that interval.
Notice that this says nothing about uniqueness. Remember the example f(x) = x
2
+1 for
x ∈ [−2, 2] had two points where the absolute maximum was obtained. Also, note that functions
that are not continuous and not deﬁned on a closed interval can still have extrema.
Example 3.3. Consider f(x) = sin(x) on (−1, 10). Recall that −1 ≤ sin(x) ≤ 1 and sin(
π
2
) = 1
while sin(

2
) = −1 and
π
2
,

2
∈ (−1, 10). So our function is bounded above by 1, bounded
below by −1 and both of these values are obtained. Thus our function has a global maximum
and global minimum on an open interval.
4
Example 3.4. Consider the following function on [−1, 1]
f(x) =

x
2
if x = 0
−3 if x = 0
This function is not continuous at 0 however it has a global minimum at 0 of −3 because at all
non-zero points, this function is strictly positive.
These examples show us that functions that do not satisfy the conditions of the extreme
value theorem can still have global extrema.
4 Local Maxima and Local Minima
The extreme value theorem gives us the justiﬁcation, but it doesn’t tell us anything about
how to ﬁnd the absolute maxima and absolute minima. To do this we need to look a little bit
more ‘locally’ at the graph. Consider the ﬁrst function we looked at today
P(c) =
−1
2000
c
4
+
1
5
c
3

51
2
c
2
+ 1150c
and let’s examine it on the interval [0, 200].
(Aside: The extreme value theorem states that this function has a global maximum and mini-
mum.) Notice that there there is something special happening at the point x = 40 and x = 100.
The function sort of has a maximum and minimum there but its not an absolute/global maxi-
mum nor is it an absolute/global minimum. Somehow though ‘locally’ around an open interval
containing these points, we are gaining information. This is encapsulated in the following deﬁ-
nition.
5
Deﬁnition 4.1. Let I be an open interval on which a function f is deﬁned and suppose that
c ∈ I. We say that c is a local maximum value of f if f(x) ≤ f(c) for all x contained in some
open interval of I. Similarly, we say that c is a local minimum value of f if f(x) ≥ f(c) for all
x contained in some open interval of I. These points together are known as local extrema.
Note. Your textbook uses any arbitrary interval, but requires c to be an interior point. What
this means is that the point c can be anything in the interval excluding any end points. To make
this easier and faster for class purposes, I have restricted our deﬁnition to an open interval.
Note. Global extrema of a function that occur on an open interval contained in our domain are
also local extrema.
Again though we have the issue of ﬁnding where these local extrema occur. However, this
issue can be resolved. Look at our local extrema. What do you notice about the function values
at the points where this occurs? Think derivative. Think horizontal line. Notice that at these
points, the derivative is exactly 0 at the points. This motives the following theorem
Theorem 4.2. (Fermat’s Theorem or Local Extreme Point Theorem) If a function
f(x) has a local minimum or maximum at the point c and f

(c) exists, then f

(c) = 0.
It seems that any time there is a local extrema, the derivative should be zero there. This
is unfortunately not quite the case.
Example 4.3. We look at f(x) = |x|. Notice that this function is not diﬀerentiable at x = 0
but since
f(x) = |x| ≥ 0 = f(0)
we see that it has a local minimum at 0 (and in fact, this is a global minimum).
While, just having a zero derivative at a local extrema is not the whole picture, it turns
out that the only case is fails is when the derivative does not exist. This is summarized by the
following deﬁnition.
Deﬁnition 4.4. A critical point is a point c in the domain of f where f

(c) = 0 or f

(c) fails
to exist.
In fact, all critical points are candidates for extrema but it is not true that all critical
points are extrema.
Example 4.5. Consider the function f(x) = x
3
. We saw before that this function has no
maximum or minimum. However f

(x) = 3x
2
and f

(0) = 3(0)
2
= 0 so the point x = 0 is a
critical point of f that is not an extrema.
With this newly created toolbox and terminology, we now discuss an algorithm for ﬁnding
extrema.
6
Algorithm 4.6. (Algorithm for ﬁnding global minima and global maxima) Let f be
a continuous function on a closed interval [a, b] (so that our algorithm satisﬁes the conditions
of the extreme value theorem).
(i) Find all the critical points of (a, b), that is, the points x ∈ (a, b) where f

(x) is not
deﬁned or where f

(x) = 0 (usually done by setting the numerator and denominator
to zero). Call these points x
1
, .., x
n
.
(ii) Evaluate f(x
1
), .., f(x
n
), f(a), f(b), that is, evaluate the function at all the critical
points found from the previous step and at the two end point values.
(iii) The largest and the smallest values found in the previous step are the global minimum
and global maximum values
Aside: There is one small issue here if we got a computer to use this algorithm. It is
possible that you could give it a function where this algorithm doesn’t terminate. As a test of
your understanding, can you come up with an example?
Let’s ﬁnish up with a couple of examples of how we would use this algorithm.
Example 4.7. Compute the absolute maximum and minimum of 3x
2
−4x + 2 on [−1, 2].
Solution: Our function is continuous (and in fact diﬀerentiable) everywhere. Hence we
may use the algorithm. We compute the derivative and see that
f

(x) = 6x −4
Setting f

(x) = 0 and solving yields
0 = f

(x) = 6x −4
⇒4 = 6x

2
3
= x
Now, we evaluate f at x =
2
3
, −1 and 2 (that is, the critical points and the end points). We get
that
f

2
3

= 3

2
3

2
−4

2
3

+ 2 =
2
3
f(−1) = 3(−1)
2
−4(−1) + 2 = 9
f(2) = 3(2)
2
−4(2) + 2 = 6
From this, we see that the absolute maximum is 9 obtained at x = −1 and the absolute minimum
is
2
3
obtained at x =
2
3
.
Example 4.8. Compute the critical points of f(x) = 5x
2
3
Solution: We compute the derivative
f

(x) =
10
3
x
−1
3
Now we check when the derivative is 0 and when it is undeﬁned. This function is never 0 but
happens to be undeﬁned at 0 which is a point in our domain. Hence the critical points are just
x = 0.
7
Example 4.9. Let ﬁnish oﬀ with our ﬁrst example. We compute the global maximum on [0, 200]
of
P(c) =
−1
2000
c
4
+
1
5
c
3

51
2
c
2
+ 1150c
The function is continuous and diﬀerentiable everywhere so we can apply the algorithm. Taking
the derivative yields
P

(c) =
−1
500
c
3
+
3
5
c
2
−51c + 1150
Setting this to zero then solving (using a computer) yields
c = 35.89038371, 94.40553426, 169.7040820
Evaluating the function at these points and the end points 0 and 200 yields
P(0) = 0
P(35.89038371) = 16843.48591
P(94.40553426) = 9860.6282
P(169.7040820) = 23545.8859
P(200) = 10000
From this we can see that the maximum occurs at c = 169.70 and gives a proﬁt of 23545.89
dollars.
5 Summary
Let’s wrap up what we have done in this lecture.
(i) Deﬁned absolute maximum and absolute minimum and gave examples of these concepts.
(ii) Stated the extreme value theorem and gave examples of when its true and examples where
the conclusion holds even when the preconditions fail.
(iii) Deﬁned local extrema and gave examples of functions that do and do not have extrema.
(iv) Deﬁned critical points and applied to ﬁnd and classify critical points of a given function.
(v) Find the absolute extrema of a given continuous function on a closed interval.
SPOILER: An example of where the algorithm above fails is any constant function deﬁned
on say [0, 1]. The derivative is 0 everywhere hence a compute will not terminate when trying to
ﬁnd all the critical points as it is the entire interval [0, 1]. Notice that in this case the maximum
and the minimum are the same value and occur everywhere on the interval.
8