Engineering Mathematics – I

Prof. Jitendra Kumar

Lecture – 02
Mean Value Theorem

Hi. So, welcome to the second lecture on Engineering Mathematics – I and today, we will
discuss Mean Value Theorems.

(Refer Slide Time: 00:23)

So, let us go through the concepts covered. So, we will discuss the Lagrange mean value
theorem; a very important concept which is the extension of the previous lecture where we
have a studied Rolle’s theorem and there is another generalized a mean value theorem or the
Cauchy mean value theorem which will be also discussed in today’s lecture.
(Refer Slide Time: 00:45)

So, let me just recall from the previous lecture. So, if a function f is continuous in a closed
interval [a , b] and differentiable in open interval (a , b) and the function value at the point a
and the point b. So, the end points of the interval is equal, then there exist a number c in the
open interval (a , b) such that the derivative vanishes at this point.

The geometrical interpretation is as clear from this figure. So, we have a function f which is
continuous and differentiable and the function value at a and b both are equal. Then, there
exist a point c here where the tangent is parallel to the x−¿ axis.

(Refer Slide Time: 01:37)

The third condition where the function was equal at the two end points is not required here.
So, it is more general and less restrictive and in that case again there exist at least one number
c in the open interval (a , b) such that this quotient here

f ( b ) −f ( a )
b−a

is equal to the derivative at a point c. So, let us first discuss the geometrical interpretation of
this Lagrange mean value theorem.

So, if we have a function which is continuous and differentiable in some interval a , b and
then let us take a look what is this quotient here. So, if you join these two points f at a and f
at b then we get this line segment. So, what is the slope of this line segment? Let us compute.

So, in this case if I draw this triangle here the height here will be f (b) because the distance
from here to this point is f (b) and the distance from this point to this point here is f (a). So,
f ( b )−f (a) is this height of this triangle and the base here is b−a , because up to this point is
b and up to this point here the co ordinate of this point is a f (a).

So, here this distance is b−a and this one is f ( b )−f (a) and. So, this quotient here f (b)
minus f (a) were b−a this perpendicular divided by the space will be the tangent of this
angle. So, basically this expression here f ( b )−f (a) were b−a is the slope of this line
segment which we have drawn by meeting these two end points of the a curve and now, what
this theorem says that this will be equal to f prime c. So, the slope at some point in the
domain a to b.

So, the geometrical meaning is that they will be at least one tangent; in this particular case we
can see these three tangents which are parallel to this line segment. So, this Lagrange mean
value theorem says that there will be at least one point where the tangent will be parallel to
this line segment joining these two points, the end points of the curve. So, as I have written
here in other words there is at least one tangent in this interval that is parallel to the line
segment that goes through the end points of the curve.
(Refer Slide Time: 04:41)

The proof is very simple if we consider this function

f ( b ) −f ( a )
ϕ ( x ) =f ( x )− [ b−a ] x.

So, if you take a close look at this function it is a basically difference of two functions f ( x)
and minus some constant times x. So, if f is continuous in the closed interval [a , b] and
differentiable in the open interval (a ,b) and x is also a function which is continuous and
differentiable in those intervals, then this difference will be also continuous in closed interval
[a ,b] and differentiable in the open interval (a ,b).

So, for the setting of this function is done because if you compute here for example, the ϕ at
the point a and ϕ at the point b then we will realize that these two values are also equal. So, ϕ
at a is nothing, but

f ( b )−f ( a )
f ( a )− [ b−a ]a

So, this if I simplify then and this will become b f ( a )−a f (a) and then minus a f ( b )−a f (a)
and divided by this b−a. So, this a f (a) will get cancelled and then we will get b f ( a )−a
f (b) over b−a and now, if I compute here f (b). So, here you have then f (b) minus this
course in f ( b )−f (a) and divided by this b−a and then here b.
So, if I simplify now this so,

b f ( b ) −a f ( b ) −b f ( b ) +b f ( a )
b−a

So, in this case this b f (b) gets cancel and we get

b f ( a ) −a f ( b )
.
b−a

In the earliest case also we got

b f ( a ) −a f ( b )
.
b−a

So, the function is taking same value at a and b and if we recall again the condition was for
Rolle’s theorem other than the continuity and differentiability that the function should be
having the same value at the two end points. So, in this case this function f satisfies all the
conditions of the Rolle’s theorem.

(Refer Slide Time: 07:35)

And therefore, we can apply the Rolle’s theorem to this function ϕ ( x). So, what will now
give us if we take the derivative here the ϕ '( x) is equal to the derivative of x minus this is a
constant. So, here f ( b )−¿ f (a) and divided by this b−¿ a and the derivative of x will be 1.
And the Rolle’s theorem says that the there will be a point where the function will be the
derivative of the function will be 0.

(Refer Slide Time: 08:15)

So, in this case now if we apply the Rolle’s theorem then ϕ '(c) will be 0 and for some c in
the interval in the open interval (a , b) and which implies precisely that this is 0 and that is the
Lagrange mean value theorem that f ' (c) will be equal to

f ( b ) −f (a)
=0.
b−a

So, the construction of this function here was important to prove the Lagrange mean value
theorem and this ϕ here satisfy all the properties of the Rolle’s theorem and we can apply the
Rolle’s theorem to this function and we got the desired result of the Lagrange mean value
theorem.
(Refer Slide Time: 08:51)

So, there is another one the generalized mean value theorem which is also called the Cauchy
mean value theorem. So, here we will consider two functions instead of one. So, if f ( x) and
g( x) are two functions continuous in closed interval [a ,b] and differentiable in open interval
(a , b) and there is another condition on g that g ' the derivative of g does not vanish
anywhere inside the interval then there exist a point c in the open interval ( a, b )such that this
Cauchy theorem

f ( b )−f (a)
g ( b )−g(a)

is equal to the ratio of the derivative of this f and g at the point c.

So, the proof is again pretty similar to the earlier proof of the Lagrange mean value theorem
and in this case we set this function or define a function in such a way that this ϕ ( x) is equal
to f ( x )−¿ f (a) minus this quotient here which will be coming in the result of this Cauchy
mean value theorem and multiplied by g ( x ) −¿ g(a). So, again the similar argument since f
and g they are continuous in closed interval and differentiable in the open interval (a , b).

Student: (Refer Time: 10:15).

So, the ϕ is also differentiable and continuous in the given intervals. Moreover if we see here
that what is the ϕ at a, that is f ( a )−¿ f (a) here this is 0 and g ( a ) −¿ g(a) is also 0. So,
everything is 0. So, the ϕ (a) is 0 and the ϕ (b) which is f ( b )−¿ f (a) and ( b )−¿ g(a). So, this
g ( b ) −¿ g(a) will get cancel with this g ( b ) −¿ g(a) and then we will get f ( b )−¿ f (a) minus
this f ( b )−¿ f (a) which is again 0.

So, in this case the ϕ (a) is 0 and ϕ (b) is 0 and ϕ satisfies all the conditions of the Rolle’s
theorem and therefore, we can apply Rolle’s theorem to this function ϕ ( x).

(Refer Slide Time: 11:15)

So, applying the Rolle’s theorem, but before that there is a point here that we have to tell that
this ϕ is well define because this g ( b ) −¿ g(a) should not go to 0; that means, g(b) should
not be equal to g(a). The question is why g(b) cannot be equal to g(a)? We have not made
such a restriction directly in the assumptions of this Cauchy mean value theorem , but again
there was an additional condition that g '( x) does not vanish anywhere inside the interval.

So, if this g(b) is equal to g(a) in this case we can again apply the Rolle’s theorem to the
function g which will say that there will be a point c in the open interval (a , b) where the
derivative will vanish. But, as per the assumption of the theorem g ' does not vanish
anywhere inside the interval. So, this cannot be equal. So, there will be never such a situation
that this g(b) will become equal to g(a) and this will become infinity.

So, the function is well defined the function is differentiable, it is continuous and g(b) ϕ (a)
is equal to ϕ (b). So, all the conditions of the Rolle’s theorem are satisfied for this function
ϕ ( x).
(Refer Slide Time: 12:49)

So, if we apply the Rolle’s theorem now, to the function then we will get exactly the result
which is given here because the ϕ '( x) will be the derivative of f and then this is a constant
here minus again this expression and the derivative of g.

(Refer Slide Time: 13:11)

So, what will be this again? So, let me just come to this point. So,

f ( b )−f ( a )
ϕ ' ( x )=f ' ( x )−
[ g ( b )−g ( a )]g ' ( x)
and the Rolle’s theorem says that at the point x is equal to c this is equal to 0.

So, what do we get then? The ϕ '(c) and divided by this g '(c) is equal to the

f ( b )−f (a)
.
g(b)−g (a)

So, that is the Cauchy mean value theorem or the generalized mean value theorem.

So, if you now we discuss the geometrical meaning or the of this Cauchy mean value
theorem. So, here now we consider this parametric curve which is given by x=g¿); g is the
function the given function there, but I have introduced this parameter t which is commonly
used for the parametric curves and the y is equal to the other function f (t ) and t varies from
a to b in this close interval.

So, this parametric curve you can trace by varying the values of t. So, if for example, t=a,
then we have here the x co-ordinate g(a) and the y co-ordinate f (a)of this point. So, this
point is g(a), f (a) and then if we vary t we will basically move on this curve we will trace
this curve and till we reach the end point here, t is equal to b which is given by g(b) , f (b).

So, now the geometrical meaning is similar to the earlier result on Lagrange mean value
theorem. So, they will if we join these two points by this line segment, then this theorem
says; so, first of all this the slope of this line segment will be given by this

f ( b )−f (a)
g ( b )−g(a)

because of the same argument as we have discussed earlier. The height will be f ( b )−f (a)
and this the base of this triangle will be g ( b ) −g (a). So, this is the slope of this line segment
and then the right hand side here says that there will be at least one point on this curve where
the tangent will be parallel to this quotient line.

So, if you take a close look this f ' (c)/ g' ( c ) is nothing, but the slope of the tangent line at
some point c, because the slope will be calculated as at some point here the dy /dx is equal to;
for the parametric curve, so, this will be dy /dt and divided by dx /dt or the y ' (t ) divided by
the x ' (t). And, this y is basically the f , so, here you have the f ' (t) over g '(t ) and this
theorem says that there will be a point somewhere in the interval. So, t is equal to c. So, we
will get this slope here of this tangent line as f 'at c divided by g' ( c ) .

(Refer Slide Time: 16:55)

And, now let me just quickly summarize at this point what we have learnt today. So, we
discuss the generalized mean value theorem which was this

'
f ( b )−f (a) f ( c )
=
g ( b )−g(a) g' ( c )
(Refer Slide Time: 17:15)

And, now what will happen if ( x )=¿ x. So, g ( x ) =¿ x meaning that here you have this g ( b ) =¿
b and this g ( a ) =¿ a and g '( x) here will be just 1. So, what do we get in this case the
Lagrange mean value theorem because that conclusion will be

f ( b ) −f (a) '
=f ( c ) .
b−a

So, in this particular case when we take g ( x ) =x we will get the Lagrange mean value
theorem.

And, what will happen to this Lagrange mean value theorem if we put ( b )=¿ f (a), the
additional condition what we have for the Rolle’s theorem. So, f ( b )−f (a) this quantity here
will become 0 and then we will get f ' ( c )=0. So, this is the generalized mean value theorem
and as a particular case if we take the function g ( x ) =x we will get the Lagrange mean value
theorem and again if we add another condition that f ( b )=¿ f (a) we will get the Rolle’s mean
value theorem which is f ' ( c )=0.
(Refer Slide Time: 18:31)

Now, we go to the we will go to this some examples the first one that using mean value
theorem we will show that this inequality cos e x−cos e y is less than equal to x− y for x, y ≤ 0.
So, first we note that when both are equal x, y are equal then naturally this cos e x−cos e was
0 and is equal to 0. So, then in equality is naturally satisfied when x and y both are same. So,
we will consider the case when they are not same.

And, now we consider the f ¿) another function cos et because clearly we can see that we
want to prove this cos e x−cos e y using mean value theorem. So, if you consider this function

f ( t )=cos e in this interval x , y and naturally we have assume that x ≠ y and now, we apply
t

the mean value theorem to this result what we will get? cos e x−cos e y x minus y is equal to
there will be some point in this interval open interval ( x , y) and the value of this quotient will
be is equal to f ' (c).

So, this is the Lagrange mean value theorem and now we will estimate this derivative because
the derivative we can compute the f ¿) is cos et . So, taking the absolute value both the sides
we get this cos e x−cos e y and this absolute value will take to the right hand side. So, the x−¿
y absolute value and the absolute value of these f ' . So, f ' is nothing, but the n −sin et × e t .
So, because of the absolute value we have not taken this −sin into consideration. So, we
have e c sin ec because this the derivative has to be evaluated at point c, ok.
Now, this implies, so, if I we take the maximum value of this expression here. So, the c
varies from x to y. So, we have taken the c from this x to y and we will take the maximum
value of this one. And, note that the c belongs to this ( x , y) open interval, so, it is basically a
negative number the c <0 because x and y both are less than equal to 0 and therefore, the c
will be strictly less than 0 in the open interval.

So, the sin is always founded by one. So, we have less than equal to one the sin function and
the e c the exponential function for this negative argument c will be always less than 1
because e 0 is 1 and o for all a negative values it takes value less than 1 for positive values it
will take more than 1.

So, this is strictly less than 1, this is less than equal to 1. So, this expression here or the
maximum value of this derivative is bounded by a strictly bounded by 1. So, we got this
inequality cos e x−cos e y is less than the absolute value of x− y which we want to prove in
this result.

(Refer Slide Time: 22:01)

The second example we will consider that this f is the differentiable function on the closed
interval −2 to 2 and such that the value is given as −2 is equal to 1, f is given as 2 as 5 and
there is another information here that f ' (x ); f ' (x ) is bounded by 1 for all values of x in this
interval −2 to 2. And, using mean value theorem we want to find the value of the function at
0.
So, if we take a look at this problem and we want to find the value of f (0), so, we need to
apply the mean value theorem or Lagrange mean value theorem in the interval −2 to 0 and 2
to 0 and then we will get some estimate on this f (0). So, if we apply the Lagrange mean
value theorem on −2 to 0 interval what we get the

f ( 0 ) −f (−2)
0−(−2)

is equal to there will exist some c 1 in the open interval −2 to 0 so that this value will be equal
to the derivative at that point c 1.

Now, this derivative here f prime c 1 is bounded by 1. So, we know the estimate of this f ' (c1 ).
This is always between −1 and 1. So, what is this expression here? The f (−2) is 1. So,

f ( 0 ) −1
2

f (0) minus 1 divided by 2 lies between −1 and 1 because this is equal to the derivative and
the derivative is bounded by less than equal to 1 the absolute value. So, this expression here
lies between −1 and 1.

Now, if we multiply this 2 to both the sides or that we can multiply 2 to this inequality here
we will get −2≤ f ( 0 ) −1≤2 and then we can add this 1 to the inequality. So, we will get here
the 3 less than equal to f (0) and less than equal to so, here minus 2 was there plus 1, so,
minus 1 and then here 2 and then plus 1 we will get 3. So, out of this inequality we will get
that −1≤ f ( 0 ) ≤3. Again if you use the Lagrange mean value theorem in the interval 0to 2; in
the interval 0 to 2 we will get

f ( 2 )−f (0)
2−0

is equal to the first derivative at some pointc 2.

So, again here the f (2) is known the f (2) is 5; so,

5−f (0)
.
2
So, what do we have here? We have f (2); f (2) is given as 5 and minus this f (0) divided by 2
and this value again is bounded by minus 1 and 1. So, we got this one here

−2≤ 5−f ( 0 ) ≤ 2.

So, this implies that this

−7 ≤−f ( 0 ) ≤−3

So, if you multiply by minus 1 here, so, the inequality will change; so,

7 ≥ f ( 0 )≥ 3

(Refer Slide Time: 25:57)

So, this inequality we will get now that f ( 0 ) ≥ 3 and ≤7 this one which says that the f ( 0 ) ≥ 3,
but ≤7. The earlier inequality says that f ( 0 ) ≤3.

So, by these two inequalities here f ( 0 ) ≤ 3 and f ( 0 ) ≥ 3 what we will conclude that f (0) has to
be 3. So, f (0) has to be 3. So, we got the value using the mean value theorem of the function
at 0 given that those derivatives and the end points value was given.
(Refer Slide Time: 26:35)

The last example here, the function f which satisfies now that the derivative is

1
5−x2

and ( 0 )=¿ 2. Now, we want to use the Lagrange mean value theorem to estimate the bounds
on f (1). So, in this case the exact value of f (1) is not possible so, we will estimate the lower
and the upper bound for f (1).

So, again if we use the Lagrange mean value theorem in the interval 0 to 1 because you want
to estimate 1. So, 0 to 1 then we will get f ( 1 )−¿ f (0) divided by this difference 1 and there
will be some point whose value will be equal to the derivative at that point. So, out of this
inequality1)

f ( 1 )−f (0) '

Now, what is the derivative? Derivative is

1
.
5−x2

' 1
So, f ( c )= .
5−c2
Now, just note that c here is between 0 and 1. So, the lower bound of this

1
5−c 2

will be obtained when we then the c this c approaches to 0; that means, this value is always
greater than 1/5 and when the c approaches to this maximum value in the interval as 1 in that
case this will become 4, and the maximum value of this 1 over 5 minus c square will be 1 by
4. So, now we know that the derivative lies between 1/5 and 1/4.

So, what is the derivative f ' (c) is here? f ( 1 )−2. So, with this we got the inequality that
f ( 1 )−2 lies between 1/5 and 1/4 and this implies. So, to we can take to the other side and
also it has to be added to the right side here. So, this we will get 11/5 less then f (1) and then
if we add 2 here this will be 9/ 4. So, we got the estimate on f (1) that it lies between 11/5
and 9/4.

(Refer Slide Time: 28:55)

So, these are the references we used here. The Piskunov, Differential and Integral Calculus
and Kreyszig, Advanced Engineering Mathematics.
(Refer Slide Time: 29:07)

So, the conclusion for today’s lecture that we have learnt the generalize mean value theorem
and as a special case when we substitute this g( x) the other function the second function as x,
we will get the Lagrange mean value theorem and if we take another assumption that f (a) is
equal to f ( v) then this will be the Rolle’s theorem which says that f ' ( c )=0, ok.

Thank you, for today’s lecture.