The Mean Value Theorem
![]()
Let f be an increasing function on
and c be a number in
(
), we have
for all x in
.
If f ' ( c ) exists then f ' ( c ) is nonnegative.
If f is a differentiable function on
, is the derivative of f always nonnegative ?
Here is another question : If a car travels 180 km in 1.5 hours, can we guarantee it will read 120 km/hr during the trip ?
At a first glance, these two questions seem to be irrelevant; however, they are a related to the following theorem :
The Mean Value Theorem
If f is continuous on
and differentiable in (
), then there is a number c in (
) such that
f ' ( c ) =
![]()
Or equivalently, f ( b ) - f ( a ) = f ' ( c ) ( x - a ).
The Mean Value Theorem says that " For a given differentiable function over a given interval, the average rate of change over the interval equals the instantaneous rate of Change at some point in that interval".
Geometrically, this is what Mean Value Theorem means :
![]()
First, let's consider a special case for the Mean Value Theorem with f ( a ) = f ( b ) = 0.
Rolle's Theorem If f is continuous on
and differentiable in
(
) such that f ( a ) = f ( b ), then there is a number c in (
) such that f '( c ) = 0.
![]()
If f is a constant then f ' ( x ) = 0, so c can be taken to be any number in (
).
Recall that if f ' ( c ) = 0 then c is a critical number of f . Hence, f ( c ) may be a local extrema.
Now it comes the question:
" If f is a inconstant function which is continuous on
and differentiable in (
) such that
f ( a ) = f ( b ),
is it true that f must have a local extrema in (
) ? "
By the extreme Value Theorem, f attains an absolute maximum and minimum values on
.
The fact that f ( a ) = f ( b ) implies one of the absolute extrema must occur at some number c in (
), that is, f ( c ) is a local extreme value of f .
Since f '( c ) exists, by Fermat's Theorem, we get f ' ( c ) = 0.
The animation below illustrates the idea how to get the Mean Value Theorem from Rolle's Theorem.
![]()
Proof of the Mean Value Theorem
Let's take a look at the following animation to see how we can pull the graph of f such that the secant line joining the two endpoints become horizontal.
![]()
![]()
Consider the function
(
)
Note that g is is continuous on
and differentiable in
(
) such that g( a ) = g ( b ). By Rolle's Theorem, we get that there is a number c in (
) such that g' ( c ) = 0. Since
g' ( x ) = f ' ( x ) -
,
we get
f ' ( c ) =
.
Note: If f does not satisfy the condition " f is continuous on
and differentiable in (
) ", then the conclusion in the Mean Value Theorem may not hold. Here are two examples :
(1) Let
on
.
It's not hard to see that f (
) = f (1), but there does not exist c in
such that f ' ( c ) = 0.
![]()
(2) Let
on
.
Then
,
whereas g' ( x ) =
and g' (1) does not exist ( g is not continuous at 1). We see that there does not exist c in
such that
g' (c) =
.
![]()
Corollary: If f , g are continuous on
and differentiable in
(
) such that f ' ( x ) = g' ( x ) for all x in (
),
then
f ( x ) = g ( x ) + C where C is a constant.
Proof It suffices to show that if f ' ( x ) = 0 for all x in (
), then
f ( x ) = C for some constant C .
For each x in (
], the Mean Value Theorem implies that there is a c in [
] such that
f ( x ) - f ( a ) = f ' ( c ) ( x - a ).
Since f ' ( x ) = 0 for all x in (
), f ( x ) = f ( a ). Therefore,
f ( x ) = f ( a ) for all x in
.
We have seen that
( arcsin( x ) + arccos( x )) = 0 for all x in
(
) and arcsin( x ) + arccos( x ) is continuous on
.
The Corollary above implies that
arcsin( x ) + arccos( x ) = C for some constant C .
Since
,
then
. Therefore,
for all x in
.
The significance of the Mean Value Theorem is that it enables us to obtain information about a function from information about its derivative. The following examples provide instances of this principle.
Example Suppose that f (0) =
and f ' ( x )
5 for all values of x .
How large can f (2) be ?
We are given that f is differentiable everywhere and therefore f is continuous everywhere. We can apply the Mean Value Theorem on
. There exists a number c that
f (2) - f (0) = f ' ( c ) (2 - 0)
So
f (2) = f (0) +2 f ' ( c )
![]()
(5) = 7
The largest possible value for f (2) is 7.
Example Determine the number of roots of
.
First we notice that
< 0 and
> 0.
By the Intermediate Value Theorem, we see that there exists a number c in
such that
.
The derivative of
is
f ' ( x ) =
.
Hence, f ' ( x ) > 3 for all x .
Using the Mean Value Theorem, we get if x > 0, then
= f ' (
) (
) for some
between 0 and x .
Hence,
f ( x ) > f (0) + f ' (
) (
) > 5 for all x > 0.
Similarly, if x <
, then
+ f ' (
) (
) for some
between
and x,
and so
< 0.
These say that f has no root outside
.
Can f have more than one roots in
? If there exists a , b in
such that
f( a ) = f( b ) = 0
By Rolle's Theorem, there exists a number c in (
) such that
f ' ( c ) = 0. But f ' ( x ) > 3 for all x, we get a contradiction.
Therefore, f has exactly one root and the root is between
and 0.
Example (Racetrack Principle) Let f and g be differentiable for all x in
and suppose
.
If f ' ( x )
g ' ( x ) for all x in
then
![]()
for all x in
.
Let
, then h is differentiable for all x in
,
and h ' ( x )
0.
For each x in (
], by applying the Mean Value Theorem to h on
, we get
+ h' (c) (
)
0.
Therefore
for all x in
.
Here are some problem to think about :
(1). If f ' ( x )
M for
, is it always true that
?
(2). Is it true that
for all x ?
(3). How do we know that
?
(4). Consider the functions
for
and
![]()
f ' ( x ) = g ' ( x ) for all
, yet f and g do not differ by a constant. Does this contradict to what we have discussed here ?