Differentiation-I
The present chapter is going to be, to a good extent, a revisit of an earlier calculus course
on derivatives. However, our emphasis will be more on the theory, rather than the techniques of
problem solving. In the physical sense, derivative is termed as the rate measurer. Geometric cally,
the derivative at a point of a curve represents the slope of the tangent to the curve at that
point. Chronologically speaking, studying curves has always been a popular topic amongst the
mathematicians. Apollonius (c.262-180 B.C.) is credited with the construction of tangents to
conic sections. His methods were similar to the ones found today in the traditional geometry
books, where a curve is treated as a locus of points satisfying some geometric conditions. R.
Descartes (1596-1650), the founder of analytical geometry and P. Fermat (1601-1665) may be
credited to have changed this viewpoint. In the new setting, the geometric problems were recast
interms of algebraic expressions. Descartes developed a method for finding tangents based upon
fitting a circle to a given point on a curve and then using the tangent to the circle as the desired
tangent to the curve. Fermat, a contemporary of Descartes, identified curves with equations
in two variables. He also developed a method for finding tangents and solving maximum and
minimum problems using a difference quotient, identical to the one that we use now-a-days t0
define derivatives, although he did not have a theory of limits. Newton was able to describe the
rates of change, called ‘fluxions’, on the intuitive idea of limits. He introduced the notation §
to denote the rate of change of y. Leibnitz, a contemporary of Newton and another great name
incalculus, developed the notation & . J. Lagrange (1763-1813) is credited with the notations
be
SL’ (),f" (@), ..., 80 commonly used today.
DEFINITION OF THE DERIVATIVE
Definition 3.1.1. Let | ¢ R bean interval and f:1 — R, and let ¢ € I. We say fis
L£a)=f(e)
differentiable or derivable atc if the limit lim exists in R.
=¢
If this limit exists, we call it the derivative of fat eand denote it by (0.
Iffis differentiable at each point of I, we say fis ifferentiable on |
Example 3.1.2.
(a) Every constant function /: R > R is differentiable on R, and /’(c) = 0 for every ce R
(b) The function /: R > R defined by /(x) = xis derivable on R with /"(c)= | forevery ¢ € R.
87DIFFERENTIATION
88
(©) The function f(x) = | x | is differentiable at every point of R except 0,
Fig. 3.1
function f defined by
Example 3.1.3. Is the
Xsin us x#0
(x)= x
0 x=0
differentiable at)?
Solution, We have
tim £4) ~ £0)
X90 yg
This limit does not exist (see Mlustrative Example
Therefore, Fis not differentiable ato,
4 of Section TL)
1
y=xsin 2
Fig. 3.2
| Remark
i i yweven,
In the definition of derivative of fat ¢, We have taken the domain to be an interval ue veal
he de’ i ken ;
me terval is not the requirement for having derivative. What we need is that the fun
an interva u "
should be defined in a neighbourhood of the point ¢, Here 1 i
i c in to be an i
‘our results would require the domain to ,
t of |
's taken as an interval as mos!89 CHAPTER 3
Inthe examples above you may notice that a function /does not possess a derivalive atc if ts
graph has a ‘sharp corner’ at (c, f(c)), Geometrically spe:
enough to have a tangent line at (c, f(c)). However, this is just an intuitive suggestion which
may help you visualize the graph. (In higher analysis, you will ind a “smooth function” as the
one which possesses derivatives of every order in its domain),
aking, the graph of fshould be smooth
. If we replace x — ¢ with / in the definition of derivative, we get the following equivalent definition
of")
So) = tim LE*H) = Se)
0h
Since, the definition of derivative involves limit, a se
-quential criterion for derivative is
naturally an expected result.
Theorem 3.1.4. Let I be an interval, f: 1 R and let c € R. Then fis derivable at c if and
onk
iFlor every sequence (x) in Lwithx,#¢ forany n, such thax, > the sequence (Z (en) = Si |
=e
is convergent with /"(c) as its limit.
Proo. This is a direct application of the sequential criterion of limit of function and is left as
an exercise.
We now show that if a function is derivable at a point, then it is necessarily continuous at
that point.
Theorem 3.1.5. If f: > R is derivable at c € 1, then fis continuous at c.
Proo. We have, for all x € I,x #0,
10) -s(0)= (2-22) 9)
As fis derivable at c, tim L22=LO exists, which we denote by /'(c).
roe xe
Hence, by algebra of limits, we have
Jim(f@)- fo) = fim. (a2 a e-9)]
xe
=f'©)-0=0
> lim f(x) =f)
> fis continuous at c,
The converse of the above result is, however, not true,
_For example, consider f (x) = | x | Although fis continuous at 0, it is not derivable thereat
ample 3.1.2.),
a, dNino,
on R but not derivable at 0 and 1. Karl Weirstrass, in 1872, gave the following example of "
4 functio
haa . 7
which is continuous everywhere but derivable nowhere on R,
DIFFERENTIATION-1 90
Remark om
The function /(x) = |x| can be used to produce functions which are continuous on R but not derivahy
at a finite or even at countably infinite number of points. For example, f(x) = |x] 4x1 lis con ©
S00) = ¥ Leos ax)
n=02”
Example 3.1.6. Show that the function f (x) = x!4, x € R is not differentiable at x = 0.
Solution. We have
fL)-f0) _xX?-0_ 1
x-0 x ae
Now, we show that lim a does not exist.
£90 x23
1 1
GB -Thenx, > 0, but (g(x,)) = (==) = (7?) is not convergent
in R. Hence, lim g(x) does not exist.
x30
That is, lim LO- LO
x30 x-0
Therefore, fis not differentiable at x = 0.
Example 3.1.7. Show that the function f defined on R
by
1
We write g(x) = ay . Let x, =
x
does not exist.
x? if xis rational,
Sx) = {
is differentiable at 0 and find f".
Solution. We have, f(0) = 0 and
LE)- FO _ fo. -{; if xis rational
0, if xis irrational.
x-0 x 0. if xis irrational
Hence, tim L-S® _
x30 x-0
> f(0)=0
Hence, fis differentiable at x = 0.
ILLUSTRATIVE EXAMPLES =——
at oan
1. Suppose fis differentiable at the point c and that fis a function continuous
t(h) = 0 only at A = 0. Prove that
S(e+t(h)) = fle) _ py
man OCHAPTER 3
1
A ji i li =1(0)=
solution. Since, ¢ is continuous at 0, pum wh) = (0) =0
h->0 => h)>0.
S(e+ th) =f) _ L(e+th))- fe) _ p,
Hence lim “~~ (py dig hy =£)
Then,
2, Letf: R > R be differentiable at c and f(c) = 0. Show that
g()=1/@9 lis differentiable at c ifand only if f"(c) = 0.
Solution. We have,
voy = tim Leth Sle)
oa ee |
= tim £+4
hao h
gle+h)~ gc) _ |f(e+h-|S)
h h
_ lee+)
A
[+ S()= 0)
Now,
Suppose, f’(c) = 0.
S(e+h)
This implies lim
nao A
0,
andhence lim (fe) =0 E lim f()=0 > lim (- Fo=0]
ao Eee fear
> g(c)=0.
Now, suppose f'(c) =/# 0
= tim LE+") _,
aso
1 1
Suppose, /> 0. Consider the sequences (« + ) and (< ~ 4)
#0
1 1
By the sequential criterion for derivative, as c+ ~ > ¢ and c— + ewe have
n
Aeeshee Aes
2 31 and ——“—_ >!
In —Vn
Ey S| e+ r(e-4)
n
ASI wv f()=0.
mot and [Ie sf
a 1 1
sfe+4) >0 and 1(e-4) <0 or agen [+ [> 0]
n n
>DIFFERENTIATION-|
He) oy ow Had
a Un
sel
ct+h ' cnet
lim [rex hl does not exist, by the Sequential criterion.
ae hoo
_ (fer Ay :
Similarly, we can show, if/7 <0, Jim, Inte) does not exist.
Hence, g(x) = |/(0] is differentiable at ¢ if and only if/"(0) = 0.
3. Letf: ROR be defined by,
x wn( 5) for x40,
f= x
0 for x =0.
where r> 0 is a rational number. Determine, for what values of 1, f'(0) exists.
Solution. We have,
a
_ sin—
lim x"! sin— = lim —*
x30 x x90 xP
, then x,, > 0 but
nv?
If we take, x, = = (20m + *) diverges to + «0. Hence,
Pq
Xn
2nt +>
2
. sin Vx .
lim —{,— does not exist.
PIG
Pay
= wm x’! sin 2 rar 7
Ifr=1,then lim x’ 'sin(+) = lim sin, which does not exist (See Illustrative Example*
x70 x x30 x
of Section 4.1).
Ifr>1, then six]
yr! sin 2
x
Now, lim (x”"!) =O asr—1>0.
x70
Hence, by the Squeeze Property, lim (» ye
a 5
Lx) - SO)
x-0
Thus, Jim, exists only for r> 1.
Therefore, /"(0) exists only for r> 1,CHAPTER 3
93
4. @ Prove that if fis differentiable atc, them ny fer
. ao
sor
( Give an example of a function fand a point ¢ such that tim Lerh)~ fle-h)
exists but fis not differentiable at c. ae 2h
((¢ —4) exists and equals
solution. (a) We have
2h
_'YLlerh-s0 , fe-)- se
2 A
Since f’(c) exists, we have
tim LEEM=LO _ jim LE*M-SO _ pe)
ho h eth>e (ct+h)-e
Le-M-SO) _ pe)
im LE-M-fO 4,
ae 1D lit. (e=hyae
Le+)=SC-) _lLinwap@er
Hence, jim AEFI ELE = “rors Ol=s'
(6) Consider the function f(x) = |x| and the point c= 0. We have
tim LOtM= SOP) < ig VEAL _ g
hoo 2h hoo 2h
But fis not differentiable at 0. (See Example 3.1.2)
EXERCISES 3.1
1. Using the definition, find the derivatives of the following functions
@se=1,x0 (6) g@)= vVx,x20
x
(Aw=-,x>0 d) h (x)= ~~ x#-1
aye aie Ore xtl
(e) k(x) = ax? + bx + ¢ (a, b, c are constants)
Olaj=23 @m@)==4 we.
2. For each of the following, determine where the function is not differentiable
(f(x) =|9-2| (6) g(x) =| 2x |-|2-x]
(h(x) =x] x] (d) k(x) =| sinx |
iG)=|0| ie)=el!
{g)1(x)=|xp
— —_—eDIFFERENTIATION-| %
3, Assuming that f"(c) exists, find the following limits
— f(c+2h- fo) im Le=M= Sle)
© fig Oe
hy- f(c-h oo
© fim Sct fe ) (a) fim, FEE) MO
4. Which of the following is differentiable at x = 0?
(a) f(x) = sin |x| (b) g (x) = cos |x|
5. For each of the following, determine the values of the constants a, b, c and d so that the
function is differentiable on R
ax+b, x<0, sinax, x<0,
(@f@=srtl OSxs2, (i) g(x= yoxte, O
2. x43, x2d.
6. Determine if the following functions are differentiable at x = 0
x, xis rational
_ fly xis rational 7
(@FC)= Jo, xisirrational 8@= 10, xisinational
ve
sin x, xis rational x5 x40
(A=) x, xisirrational @iw=y rl
0, x=0
: x sin Z x#0
@I@&= x
0, = x=0
Suppose that fis differentiable at c and f’ (c) # 0. Show that there exists 5 > 0, such that
O<|x-cl S@)4#f(O).
8. Suppose f: (a, 6) > R is continuous on (a, 6) and differentiable at c € (a, b).
»
$O-S0 4,
Define g:(a,b)>Rby —_g(x)= x-¢ >
f(x, x=0
Prove that g is continuous on (a, b).
Prove that iff: R — Ris a differentiable even function, then f” is an odd function.
(If f(-x) = f(x) for all x, then fis called even, if f(x) = -f(x) for all x, then fis called ol!
. Prove that if g : R — R is a differentiable odd function, then g’ is an even function.
°
Ss
. Show that if fis differentiable at a point c, then
S'(0)= sim (o[(e+4)- ro)}CHAPTER 3
95
How ,
the existence Off (eo).
R > R be defined by f(x) ~ x? sin(Le’) fora 4 0 and £(0) = 0. Shaw that fis
“viable for all x © R. Also show that the derivative /’ in not bounded on {4,1
ever show by un example that the existence of the limit of this sequence does not imply
2, Let f
ittere
RULES FOR DIFFERENTIATION
high school calculus, students are made familiar with the rules for differentiating the
In the a
quotients and composition of functions etc, In this section, we revisit them and
ums, products,
‘eview the theory associated with their formulas.
fe
Theorem 3.2.1. (Power Rule). For any given natural number n, the function f(x) = 2" is
Giferentiable everywhere on Rand for any © € Ry/"(e) ~ net
proo, For n= 1,/(x) = x 80 that /"(c) }= Le! forevery ce R.
Hence, we assume n 2 2. Then for any ¢ © Randx vc,
_ (xn eye 4x" et tact +e!)
xe
axle tet ter +
Hence,
= lim (x +x'e+.., bxet +e!)
N+ tet, +cctt+ on!
d+ ol +, +e "(n — times)
=nc"™ J
Therefore, f(x) =x" is differentiable on R and f’(c) = nc* for any c € R
Theorem 3.2.2. (Algebra of Derivatives) Let I< R be an interval andc € 1. Letf, g: 1>R
be differentiable at c.
(a) Ifa € R, then the function af is differentiable at c and (af)(c) = af'(c);
() {+ gis differentiable at c and (f+ g)'(c) =f'(c) + 8();
(©) f- g is differentiable at c and (f- g)'(c) =/"(e) - £0
(4) (Product Rule) fg is differentiable at c and (fg)'(c) =f"(o)g(©) + S(O’
(e) (Ouos; . . eee .
© (Quotient Rule) If g (c) # 0, then the function Lig differentiable at ¢ and
&
(fJo _ Logo - SOK")
x (g(o))”DIFFERENTIATION-1
96
Proof. We prove (b), (d) and (e), leaving (a) and (c) as exercises for the reader
(6) im yoe tim, ve
= tim LED= LO) + 20) glo)
xe x-¢
= tim LOL ae LC) jig SOD = (0)
xe re xe
=f") +g ‘O
Hence, (f+ g)' (c)=f'(c) +g’ (0)
@ tim Me Ueto = lim
LO g(X) - (80) + (g(x) - f(g(c)
= lim
re xe
= “ LO geys fle A= =8)
= tin L=LO Le
«fim gto) + f(e). lim
=f" (xo flo’ (o)
as lim g(x) = g (c), because g being diflerentiable at c, is continuous at c.
xe
Hence, (72)' ()=S'(c) g (c) +f(c) ¢ (0)
(©) Since, g (c) #0 and g is differentiable and hence continuous at c, we get g(x) #0 on some
neighbourhood V of c. Forx € V7 I, we have
Lay-Lig L0_L0
g & _ ax) gle)
xm xm
- ee = fogs)
z-e &(x)g(c)
g(x) ~ gc)
x-c
wag | ee Sge)+ fOg(- Lroto|
ae x-¢
L0)= SO), g@)- se-«09|
awl eae 8) S(O.
Now, lin 5 “30 as gis continuous at c and g(x) #0 in VAL
Hence,
Log_f
“(x)-*-(¢)
-10]
lim £—_& = tim L2)-S) LO. 6) se). lim oe
roe xme war ee
xe xeCHAPTERS
97
lien ore) fle) elle
aaa [/').g() - S00. 8'(C)]
Therefore,
(4) (¢ = LOK = LOK)
8 (g()?
1 F . a
The formula for (2) (0) is obtained from (e) by taking /to be the constant function with
f(x) = | foralll x.
We have
( 1 } (= L8O=10) ___ #0)
g (g(o” (go?
using the fact that derivative of a constant function is zero.
Now, we turn towards proving one of the most usefull results in differentiation, called the
chain rule. It provides a formula for finding the derivative of a composite function fog in terms
of the derivatives of fand g. In the elementary calculus, students have already encountered it in
the familiar form: Iy is a differentiable function of t and 1 is a differentiable function of x, then
od a
de dt de
First we prove the following result which will be used in the proof of the chain rule.
Theorem 3.2.3. (Caratheodory’s Lemma). Let f be defined on an interval I and let ¢ € I.
Then, fis differentiable at c if and only if there exists a function f;: I > R such that f is
continuous at c and satisfies the relation
S)-f()=@-OF@), — forallx eI.
Moreover, if the conditions hold, then f"(c) =fi(c).
Proof. Let fbe differentiable at c. We define f;: 1 R by
LO-SO ae
f@=4 xe .
fd. xe
Then jim f(x) = tim LO=~LO = 7(e) as fis differentiable at c, That is,
rae roe X=€
jim f@)=f(0).
Hence, fis continuous at c.
From the construction of f, it follows that
L(x) ~f(0) = (x - fi (x) forall x € Lx 4.
Hwe put x = ¢ in (i) then both [Link]. and RIS. become 0, Thus, x
Conversely, let the conditions hold. Then, by the continuity off) at ¢, we get
ESDIFFERENTIATION-1 98
lim f) = fi(c)
= tin LL. — fey
=> fis differentiable at ¢ and (“(c) = fi(c).
Remark =
‘The function, fying (i) (8) =e) = (8 ~ &) fi (8) for all x € 1, and (i) fis continuous at
lied the increment fimetion associated with f and c, Existence of increment function ensures the
rentiability of fat ¢; also the derivative of fat ¢ is the value of the increment function at c
di
Now, we move towards proving the Chain Rule. Iff is differentiable at cand g is differentiable
‘/(c)s then the Chain Rule states that
(go! (8 SOLO)
Now for /(x) #/(c), we may write
(gof ) = (8) _ 8) - 8UO) SO)- FO
xe FO-f x-e By
Since, x > c implies f(x) > f(c) (F being derivable, is continuous at c), we should get
(i) from (ii) by taking limit as x > c. Indeed, this would be true provided f(x) # f(c) in some
neighbourhood of c. For example, if fis strictly monotone, then this holds. However, in genera
Wwe cannot assume that f(x) #/(c) in some neighbourhood of c. Hence, we have to adopt some
other method for proving (i). The Caratheodery’s Lemma comes to our help in this regard,
fi)
‘Theorem 3.2.4. (Chain Rule). Let I, J be intervals, f: I» R and g : J -» R be such that (l)c
Jandc € L. If fis differentiable at c and g is differentiable at f (c), then the composition
function gof'is differentiable at c and
(gof) (= FO).S'O):
Proo. Since, fand g are differentiable at c and f(c) respectively, by Caratheodory's Lemma
there exist fj:1 > Rand g;: J > R such that
OS) -SO = &- off), and
8) - 8) = 0 -FO)8i)
(ii) fis continuous at c, g; is continuous at f(c)
and fc) =f"); 8) = 8 O)
On taking y = /(x), we get from (i).
af) -8 FO) = F@)-SO) gf)
= («OF @) gi FQ)
= (xc) fi) (8 of) &)
= (SO) - 8S) _ (eo A(x)
x-c
aa CHAPTER 3
As fis continuous at c and g, is continuous at /(c), g, of is continuous at c.
Hence, by taking limits on both sides as x > c, we get
jim £LEI= RCC)
x
xe
= lim (g, of) (x). lim (f(x)
= (g)0f) (c).f (0)
=a S(o)) file)
=P SO)S() (from (ii)
= lim
sim =F SOL"
Hence, gofis differentiable at ¢ and (gof)'(c) = ¢'(f (O(c).
While solving problems, usually the process of differentiation of a function and that of its
inverse are carried out independently of each other. Below we show that the two can be linked
(gof) (x) - (gof) (c)
x-
inanice way.
Theorem 3.2.5. (Inverse Function Theorem for Derivatives) Let fbe a continuous, strictly
monotone function defined on an interval I and let be derivable at c € I and f“(c) # 0. Then
f-lis derivable at f(c) and
1
ly, —
FY FO) FO
Proof. We shall use the Scquential Criterion for differentiability (Theorem 3.1.4.). Let (¥,) be
a sequence in f(I), converging to f(c) such that y,, #f(c) for any n. For each n € N, let x, = cam
\y,). Then x, # ¢ for any n € N. Now, f being strictly monotone and continuous on I, f~! is
also continuous (Illustrative Example 5 of Section 2.3). Hence,
Yn >F(c) implies f(y) > (F(0)). Then
— (£7 O)-f UO )_ ( mre )
in| y fo |) Fe)-F0
= lim ———_-_~
(22)
Xe
[By algebra of limits of sequence]
My ne
1
——
im( £0 =f)
1 py the Sequential Criterion for differentiability]
Se)
Hence, by the Sequential Criterion for differentiability, fis derivable at/() and (fy (Ke)DIFFERENTIATION-1 100
Example 3.2.6, Consider the function /(x) = x? in the interval [0, +).
‘The function fis continuous and strictly monotone on (0,
given by f'(x) =.[Link] (fof) (x) =x. Hence, by the Invers
I
). The inverse function ¢
in
unction Theorem
UY Ged = Faqs Provided * (x) #0
That is,
0) @)= x40
or WO) Pa fiey> 06 =x7,x#0)
Example 3.2.7. Consider the function f(x) = 2x + 3. The inverse function is given by f-1
V=3 eon
F 88 (0f-")(x) = x. The derivative of ft is given by
(x)=
rey a!
PY IN= FeH73
> )0)= ; for each y € R.
Remark
If Tis any interval and f:1—> R is 1-1 and continuous, then it can be shown that fs stily
monotone. (The proof is a little tricky and hence avoided in this book). Therefore, Theorem 3.25 can
be reframed into the following :
Let f be a continuous one-to-one function defined on an interval I. Let f be differentiable at ¢ 1
and f"(c) # 0. Then fis differentiable at fc) and
1
“Gey =
YCQ)==.
TYVO= FG
Example 3.2.8. For n € N, the function x! is differentiable everywhere on its domain, ex!
1
d Let
at Oand (xl) = =x"
Solution. The function /(x) = x" is continuous and one-to-one (the domain of fis tobe restric)
to [0, + 2) ifm is even, to maintain one-oneness). Hence, we have, for x #0
1
FY F@)= Fw * Provided s" (x) #0
1
> OY @)= Tr #0
1
Vint =
)
> vonCHAPTER 3
404
1
x40
2
'
>
\ 1
Alen) Le areo
> dx n
ILLUSTRATIVE EXAMPLES
1, Given that the function f (x) = x3 + 3x +2 has an inverse 1 on R, find the value of
. (7) (at the points corresponding to.x= 0, 1, 2.
Solution. We have, for /"(x) # 0, x € R,
1
F'Y F@)= =
Now,
VHSB 43x42
and IQ) = 3 +3
Therefore,
x=0 => y=f(x)=2andf' (x)=3
x=1 => y=f(x)=6and/" (x) =6, and
x=2 => y=f(x)=16andf" (x)= 15
Hence,
1 1 1
iy ya bt cpap 2 “yy 16y= &
CY R= 5.0 = | andy 6) = =
2. Given that f(x) = sin x, x ¢ [-1/2, 1/2] and f (x) = cosx, find the derivative of fx)
forx © (-1, 1). Why is f~! not derivable at x=—1 and = 1?
Solution. /(x) = sin x is continuous and strictly increasing in the interval [-n/2, n/2]. Hence, by
the Inverse Function Theorem for Derivatives, (f-1Y exists at the points x where /’ (x) # 0 and
1
'Y F@)= =~
: I(x)
> 1
CY (sinx) =
cosx
> y i 1
YO)=
vi-y?
>
SY (=
>DIFFERENTIATION-I
‘
102
Now, f” (x) = cosx = 0 at x = 0/2, -n/2.
Hence, f~! is not derivable at f(x/2) and f(—n/2).
That is, f~! is not derivable at 1 and -1.
2.
Gi) D'gy= "of 9, where f =
ro
EXERCISES 3.2
The second derivative of f, denoted by f” or f*) or D?(/) is defined by f” = (f")’,
Similarly," =f =(f") and so on. If fand g have derivatives of all orders, that is
2”) exist for all n € N, then show that
OF g)" =f". gt fig’ + fer
GF BY" =f". B+ Sf gt Se" + fg”
f.@=g
(Use Mathematical Induction for (iif), which is known as Leibnitz’s formula)
3.
a
x
Suppose f and g are differentiable in a neighbourhood of the point 2 and that (f )(x) =
2X (2) = 4," (2) = 2. Determine g(2) and g'(2).
x 1
. Suppose fand g are differentiable functions f(g(x)) = 2x? + 6x + 5, g(1) = 4, g'(1) =5. Find
S(4) and f" (4).
. Suppose h, k are differentiable at 0, 0 < h(x) < W/2, 0 < k (x) < W/2 for all x and A, k satisly
the given equations
sin(i(x)) = cos*x + x + xh(x) — 1
(sin x) = cos*x + x + xk(x),
Find h'(0) and 4’(0).
. =)
. Given that f(x) = cosx, x € (0, x] is strictly decreasing cos 0 = 1, cos m= —1. Show that
is differentiable on (-1, 1), but not at -1 and 1.
- Given that g (x) = tanx, x © (-n/2, 1/2) is strictly increasing. Show that g~! is differentiab#
and D(g'0)) = (1 +37)" fory eR.
dk
. Prove that if fis differentiable at xp and f(x) + 0, then <( 1 )- = Lo),
S0)) fo)
. : s se
} Show that the function /(x) = 2x3 + 3x + 5, defined on R, has an inverse. Using the “I
ve pee : g thes
Function Theorem for Derivative’, find the derivative of its inverse function at P®
corresponding to x=0, 1, 2.
q
. Suppose fis differentiable ata, gis differentiable at /(a), and his differentiable 2! oe ie
State the Chain Rule for (hogof)(a). Apply it to find the derivative of sin? + !)
result D sin x = cos x)CHAPTER 3.
103 (x? + and let g be its inverse, Compute the following
sere
ag and g'(-3) (4) g'(42) and g'(-42)
og
PROPERTIES OF DIFFERENTIABLE FUNCTIONS
ike comity, differentability ofa function permits us to draw various conclusions
« nction. Whether a function is increasing or decreasing, where the function
timer a given function is derivative of another function ete, ean be eastl
saves possessed by differentiable functions,
about the
attains its extrema,
'Y ascertained using
erties
properties
Interior Extremum Theorem
Definition 3.3.1. Let 1c R be any interval and f: 1» R. Then
(a) fis said to have a local maximum or a relative maximum at ¢ € I, if there exists a
S-neighbourhood Va(c) of ¢ for some 8 > 0 such that f(c) > f(x) for all x € Vg(c) AL.
(b) fis said to have a local minimum or a relative minimum atc e€ I, if there exists a
S-neighbouthood V9(c) of ¢ for some 8 > 0 such that f(c) < f(x) for all x € Ve(c) AL.
We say, fhas a local extremum or a relative extremum at c if it has a local minimum or local
maximum at c,
Theorem 3.3.2. (Interior Extremum Theorem). Let I be any interval andf: 1 R. Let c be
aninterior point of I such that Ff has a local extremum at c. Then, if, “f'(c) exists, it must be zero.
Proa Suppose that {has a local maximum at c and that f"(c) exists. Then, we have
fim, LO=LO _ pe) jim L0=L0
rae’ oxae rag xne
Since, fhas a local maximum at c, we have for points near c, f(x) —f(c) < 0.
Therefore, when x > ¢, forx > c, we have
L-S0 ¢4
x-¢
. lim LO-SO <9,
xer x-c
And when x5 ¢ for x < c, we have
L0)- SO, 4
x-c
> . +
tim LO=-LO)
> xe x- -
> Sc) < Vand f(c) > 0
Sc)=0
Proof f
Of for local minim
um is similar and is left as an exercise.FFERENTIATION-1 104
Remark
(i) The converse of th
differentiable at every point in Rand /"(0) ~
minimum at x = 0.
e derivative of fmay fail o exist at a point of local extremum. For example, /(x) ~ |) hy,
ever fis not differentiable at x = 0.
ove result necd not be true, For example, the funetion f(r) —
+ why
0. But /has neither a local maximum nor a tog.
a local
(i) Th
a minimum at x = 0, how:
Wwe have proved that a function, continuous on a closed and boundey
In the previous chapter,
nd minimum values. In view of the Interior Extremum
interval [a, J, atlains its maximum al
Theorem, we have the following result for such functions.
3. Suppose fis continuous on [a,b]. Then the maximum and the minimum
Corollary
values of fon [a, b] must be attained at either
(a) the end points a or b of [a, 6], or
(b) points c € (a, b) at which either f"(c) = 0 or f"(c) does not exist.
Example 3.3.4. The function f(x) = 2x? ~ 3x? + 4 in [-2, 2] is continuous everywhere and hence
attains its maximum and minimum values at [-2, 2].
‘The derivative f"(x) = 6x2 — 6x = 6x(x — 1) vanishes atx = 0 and.x = 1. Hence, there are four
possible candidates at which the extreme values of f can be attained. They are 0, 1, -2 and 2.
Now, f(0) = 4, f(1) = 3,,f(-2) = -24 and f(2) = 8,
Hence, the minimum value of fis —24, attained at x=
2; the maximum value of fis 8, attained
atx=2.
Example 3.3.5. Consider the function f(x) = [x?— 1| on the interval (0, 2]. We have
fe) =x-1 1
: CHAPTER 3
.
ma 3.3.6. Let 1 R be an interval, /: I> R and ¢ be an interior point. Suppose that/is
” 0, er
viterentiabl atc. Then
@itso? then there exists 8 > 0 such that f(x) > /(c) forall x € I withe 0 such that /(x) > f(c) for all x € | withe -8 R has a limit lat ¢ € I such that I> 0, then there exists 8 > 0 such that g (x) > 0 for
alixel withe-8 0 such that for all x © I and c-80
x—¢
> f@)-f9= F2=LO eg s6 forall x € I,c @)>S(Q) for allx ele 0. If f'(@ < k 0 such that g (x) > g (a) for all x € [a, b] witha k > f(b), then we take g(x) = fx) — ke
Seed as above,
a Prove that there is no differentiable function f, defined on any open interval 1
' 8uch that for all x € I,
1, x20
f@)= {i x<0
PeDIFFERENTIATION-|
Solution. We assume that such a function fexists, Letahe}
differentiable on [a, 6] such that f"(a) = 2 and f"(b) =
108,
with a <
1. By D,
Wween f(a) and f(b
Vand ho 9 4 he
arboux's Theorem, there
© € (a,b) such that f"(c) = 3 as ; lies bet
- But this ¢:
Such a function cannot exist,
The Mean Value Theorem
in Value Theorem, we first Prove the followin,
as Rolle’s Theorem,
Theorem 3.3.9 (Rolle’s Theorem) Suppose that Fis continuous on the closed interval la,
and differentiable in the Open interval (a, b) and that f(a) =f(b). Then there exists at least
Point c € (a, 5) such that f"(c) = 0, .
Proa. Since, fis Continuous on the
Theorem, there exists c,
inf {(/(@):xe [a, b]}} =m (say).
Now, ifM =m, then fis a constant function and hence f"(x) = 0 for all x e[
theorem holds trivially.
ig result, commonly 0
closed and bounded interval
[, 5], by them:
d [a, b] such that if (c)
aximum Minimum
= sup tf) :x € [a, b)}
=M (say) and f (a=
, 5). Thus the
IfM m, then clearly both M and m cannot be equal to f(a),
Suppose, M # f(a) (and hence M F(b), as f(a)
Hence, fis differentiable at c. Accordingly,
Similarly, if m # f(a), we have fd) =0.
Geometrically,
(a, b) and f(a) =f
=F(b). Then we must have ¢ ¢ (6,
by the Interior Extremum Theorem, F'()=0.
Fig. 3.3
Suppose, further de iti ’s Theorem, thet
ler the conditions of Rolle’s pa
further, that a and b are zeroes of f, Un i seen
i ’ 5 5) such that /"(c) = 0. Hence, algebraically, Rolle’s Theorem implies that bet"
exists c € (a, ;
two zeros of f, there exists at least one zero of f’.
k
con xt - 3K
le 3.3.10. Show that there is no real number k for which the equation x
Example 3.3.10. :
0 has two distinct roots in [0, 1].View
. 7
. CHAPTER 3
40 ‘ —
son, Irpossible, Iet there exist some value of & for which x” — 3x + & = 0 has two distinet
earn and q in [0,1] where p ) AIO /(9) = £4) =0.
D :
ote by Rolle’s Theorem, there exists ¢ € (p, q) such that /"(c) = 0. This means
lence.
3c?
5 Bute € (p,q) <0, 1].
Thus, we get a contradiction. Hence, our assumption is wrong.
dAsaconsequence of Rolle s Theorem, we obtain the ‘Mean Value Theorem’, also known as
‘tagrange’s Mean Value Theorem’.
Theorem 3.3.11. (The Mean Value Theorem). Suppose that f is continuous on a closed
interval I = [a, b] and differentiable in the open interval (a, b). Then there exists at least one
point ¢ € (a, 6) such that
f(b) f(a) =f" - 4).
Proof. Define g (x) = f(x) —/(a) — s (x -a) , where s
S(b) = f(a)
b-a
Ai)
Then g is continuous on [a, 5], differentiable on (a, b)
and aa =sa-s(a- L9=L® (4a)=0,
ya
and a) =/0)- fa - LO LO b-a=0.
yaa
Thus, g satisfies all the conditions of Rolle’s Theorem. Therefore, there exists ce (a, b) such
that g (c) = 0.
Now, from (i)
reo=po-LO=L@
Hence, wehave —f'(c)- LO=-S@ _ 9
b-a
or
Geometrical LO)-fO=f Ob-a) .
Such tha ie ly, the Mean Value Theorem states that there is at least one point c between a and b
the point tangent line at (c,f(c)) on the curve y=/(x) is parallel to the line segment through
S (4, f(a) and (6, f(b)), if fis smooth (derivable) between a and b.
Slope = f(c)DIFFERENTIATION-1
108
The motivation for choosing the function g is also suggested from the geometry
Suppose, y = L(x) represents the line joining (a, f(a) and (b, f(b). Then La)
£00) = 1a) (
hoa
satisfies the conditions of Rolle’s Theorem. Therefore, we get ¢ € (a, b) with #(c)
May.)
J (b) and L'(x) = slope of the line = = 5). Hence, the function (0) = f(xy
Nr)
Ary
Me) 1%
LO) One
a
is f' (ce) = LE
= 0, that is, 7" (c) = LE.
hat is, f" (c)
Another alternative proof is suggested in the Exercise
may verily that L(x) = f(a) + s(x - a).
(See Q 13, Exercises 6.3).
Remark —
|
If we write b = a + A, then the conclusion of the Mean Value Theorem can be written as
z =F (a) + hf'(a +0 h) for some 0 € (0, 1). "|
-12. Verify the Mean Value Theorem for the function f (x)
the interval [0, 2].
6x? — 2 in
Solution. Since, fis continuous and derivable (being a polynomial),
we should get one point
cin (0, 2) such that
rye) = L2)= £@) _-18+2 _
fe= =u
Now, FX) = 3x2 = 12e.
Therefore, ¢ must satisfy
3e -12e = -8
> 3e2- 12e+8=0
2 2
> en 2#V8 5425
6 3
2
Hence, e=2-2 3 as2+ 3 V3 ¢ (0,2).
Example 3.3.13. Apply the Mean Value Theorem to show that
|sinx—sin yl <|x—y | forally,y eR
Hence, show that ~x $sinx <
Solution. The function f(x) = sin x is continuous and derivable in whole of R. Hence for any
y € Rwe get one c lying between x and y such that
SL) -LO)=F'(O) (ey)
sin x —sin y= cos c(x—y)
u
> |sinx—sin y|=|cosc||x-y|
<|x-ylas|cose|< 1.
Hence, |sinx-siny|<|x—y| forallx,y © R.109
CHAPTER 3
Now, taking y= 0. we Bet
|sinx-0|<|x-0|
|sinx|<|x|forallxeR
2
ipfurther, x20, we get —x Ssinx 49 7< ve <
8
1011
> <<<
8 ve 7
Hence, 1 Gi-7<1
8 7
1 1
a 7+ -DIFFERENTIATION
119
Since, f’(c) = 0, we have f(y) > f(x). Since, x, y are arbitrary points of [a,b], we a
that fis increasing on [a, 5]. le
Conversely, let fbe an increasing function on [a, 5]. Then for any c (a,
5), F(x) =Ooit
x>cand f(x) so20
Since, c is any arbitrary point of (a, b), we conclude that f"(x) 2 0 for every x € (a, 5),
(b) Can be proved similarly.
Theorem 3.3.16. Suppose fis continuous on [a, b] and differentiable on (a, 5). Then, we
have the following
(a) If f(x) = 0 for all x € (a, b), then fis a constant function.
(b) Iff"(x) #0 for all x € (a, b), then fis 1-1.
(©) Iff"(&) > 0 for all x € (a, b), then fis strictly increasing on a, b].
(d) Iff"(x) <0 for all x € (a, 6), then fis strictly decreasing on [a, 5].
Proof.
(@) Let x € [a,b], x # a. We apply the Mean Value Theorem on the closed interval [a,x] t
obtain c € (a, x) such that
SL) -S(@) = fe - a).
Since, f’(c)= 0 by the given condition, we get f(x) = f(a). Since, x is any arbitrary element
of [a, b], x # a, we have f(x) = f(a) for every x € [a, b]. Hence, fis a constant function.
(b) Let x, y € [a, b], x #y. If x y, we consider the interval [y, x] and proceed similarly to show that f(x) #/0)
Hence, fis a one-to-one function.
tain
(c) Suppose x, y € [a, b] and x < y. We apply the Mean Value Theorem on [x, y] 0°
¢ (x, y) such that
f0)-f@O=F'(OW-x).
Since, f"(c) > 0 by the given hypothesis, we have f(y) >()-
Hence, fis strictly increasing on [a, 5].
(@) Similar to (c).23
wt CHAPTER 3
corollary 33.17. Suppose /and g are continuous on {a, 5], differentiable on (a, 6) and that f
rey 0 or all x € (a, 5). Then there exists a real number k such that f(x) = g (x) + & for all
re [a5
proo, consider h(x)=f(«)-g (x) forall x € [a, 6]. Then / is continuous on (a, 5), differentiable
on(a, 6) and
A(x) =f'(x) - g(x) = 0 for all x € (a, b).
Hence, by Theorem 6.3.17(a), A is a constant function.
Therefore, there exists some real number & such that
h(xy=k for all x € [a, }].
Hence, S@)=g@) tk for all x € [a, 6].
Remark.
The above theorem shows that if f"(x) > 0 for all x € I, then fis strictly increasing. The converse
js, however, not true. For example, f (x) = x jis strictly increasing on R, but f’(x) = 0 at x = 0. Similar
is the case for strictly decreasing functions. However, we have, /"(x) 2 0 for all x € I if and only if fis,
increasing on I. Similarly, /"(x) < 0 for all x e 1 ifand only if fis decreasing on I.
inthe domain R - {-1, 1}. Here fis
differentiable on its domain and f"(x) = e =)”
> f'e)>0 for |x|> V3
f'e<0 for|x|< V3.
Hence, by Theorem 3.3.17, fis strictly increasing on (a, ~J/3] and [3 , bjfor any
a<-J3 and b> V5 . In other words, fis strictly increasing in the intervals (-c, ~J3 ] and
{§,~). Similarly fis strictly decreasing on [-J3.-1),(-1,1) and on (1, 3]. (Note that
~1 and 1 do not belong to the domain of f.)
Example 3.3.19. Show that the function f(x) = 3x - 2x3 + 12x—8 is I-1 on R.
Solution. f(x) = 35-203 + 12x-8
S'(x) = 15x4 — 6x? + 12
= 3(5x4- 2x? + 4)
= 3(5y?-2y + 4), Y=
_ The polynomial 5)? — 2y + 4 does not have a zero in R as its discriminant D = (-2)°— 4.5.4
*~76 is negative,
H
nce, f40 — forallx eR.
q
hetefore, by Theorem 3.3.17, is 1-1 on R.
rpDIFFERENTIATION-
112
The Derivative Tests for Extrema
Another important application of the Mean Value Theorem is the so called “Derivative Teas
for extrema. ts
Theorem 3.3.20. (The First Derivative Test) Let / be continuous on {a, b] and diffe
in (a, b), except possibly at c, where c is an interior point of [a, 5).
Mabe
(a) If there exists a neighbourhood V s(c) ¢ [a, 5] of ¢ for some 5 > 0 such that
S'@)2Z0fore-S 0 such that
S'(x) <0 for c-8 f(x).
Consequently, S(0) =f) for each x € (c-8, ¢)
Similarly, it can be shown that f(c) > f(x) for all x € (c, c+ 8).
Then, SOZS() for allx € (x8, ct8).
Hence, fhas a relative maximum at c.
(b) The proof is similar to that of (a).
Example 3.3.21. Consider f (x) = |x | defined on R. Although fis not differentiable atx = 0, f
“(x) exists for all x © R, x # 0. Consider the interval [-1, 1].
Here S'@)=-1<0 forx<0
and S'W=1>0 forx>0.
Hence, by the First Derivative Test, fhas a relative minimum at x = 0.
Example 3.3.22. Find the points of relative extrema, and the intervals on which the
Jollowing function is increasing and those on which it is decreasing.
S(@®) =xt Z x#0.
x
Solution. S@)=x+ : x#0
roni-+
z
S)=0 > x=-1,1
For x<-l,f(x)>0
For -ILS'(x)>0.413 CHAPTER 3
‘Therefore, by the first derivative test, has a relative maximum atx = —1 and a relative
minimum at.
‘Also, fis (strictly) increasing in (— co, —1] and in [1, + c0) and (strictly) decreasing in [-1,0)
and (0, 1.
Lemma 3.3.6. along with the Mean Value Theorem, helps us devise another test for finding
the points where the extreme value of a function occurs. Since second order derivatives are
used in this test, we call it “the Second Derivative Test for Extrema”. This test is widely used
for finding maxima and minima.
in calculus f
Theorem 3.3.23 (The Second Derivative Test). Let 1 ¢ R be an interval, /: 1 Rand c € 1.
Suppose that fis differentiable on a neighbourhood Vg(c) ¢ I of ¢ and /"(c) = 0. ILf"(c) exists
and is nonzero, then
(a) fhas a local maximum at ¢ if f"(c) <0
(b) fhas a local minimum at ¢ if f"(c) > 0.
Proof. (a) Since f” is differentiable at c, and f”(c) < 0, we may apply Lemma 3.3.6 on f‘ to
get some 6 > 0 such that
S'@)>f(0) fore-80 for all x with c-8 0 and
vp deine 9, 2Inx=3
£@)= a f")=
Now, f@=0 > 1-Inx=0 => x=e
2-3. 1
“ x=6, f"Q=f"O= FZ =- GX.
Hence, /has a local maximum at x =e.
Further SX) <0 forx>e
; S'(x) > 0 forx R be continuous on [a, 4] and differentiable in (a, 5). Show tary
if
s, then f'(a) exists and they are equal.
Solution. Leth > Oanda + 0 then ¢), > a.
Hence, we have
f(at+h)- fla)
lim === = lim (cq)
hwo h novo
= lim f(c,)
lim f(x)
=> S'(a)= lim f(x)
2. Let I be an interval and let f: 1 > R be differentiable on I. Show that if f’(x) ¢0 for
every x € I, then f"(x) > 0 for all x € Tor f"(x) <0 for all x € I.
Solution.
ppose, if possible, there are xy, x3 in | such that f"(x,) > 0 and f"(x3) < 0. Then, by
orem, there exists one ¢ lying between x, and xy such that /"(c) = 0. This gives
us a contradiction. Hence, either /"(x) > 0 for all x € Lor f"(x) <0 for all.x € 1.
Darboux’s
3. Iffis differentiable on [a, 5], f(a) = f(b) = 0 and f(a) and f(b) are of the same sign,
then prove that fmust vanish at least once in (a, 5).
Solution. We are given f(a) = f(b) = 0, f"(a) and f(b) are of same sign. We take f(a) > 0.
7(b)> 0.
S'(a) > 0 = there exists 8; > 0 such that ((x) > f(a) = 0 for all x € (a, a+8))
=> Foranyc € (a,a+8,),f(c) > 0.
Similarly, /"(b) > 0 => there exists 3) > 0 such that f(x) < f(b) = 0 for all x € (b- 83,5)
=> Foranyd € (b-8, b), fd) <0.
Hence, by the Location of Roots Theorem, there
sts p € (c,d) < (a, b) such that f(p)="
Similarly, we can proceed if we take f(a) < 0 and f"(b) < 0.
4. Prove that if ag, a), ..., 4, are real numbers such that
a .4 an
og Mg Mn
n
+a,
nt
then there exists at least one real number x between 0 and I such that
ax" tay! +. +a, =0415 CHAPTER 3
solution. Define f: [0,1] > R by
4
ont a,
f(xy= yt My mol oy
J a Te tte hae.
since fis polynomial in x, therefore fis continuous on [0, 1] and differentiable in (0, 1).
_ _ % .4, 4,-
Moreover, (0) = 0 and f(1) = pelt tet on + a, = 0 (given)
rence, by Rolle’s Theorem, there exists some x € (0, 1) such that /"(x) = 0.
> at tax +... tay x ta, =0
5, Let fbe defined and continuous on [a —h, a + h| and differentiable in (a —h, a + h).
Prove that there is a real number 0 and a real number ¢ between 0 and 1 such that
(a) fat h)- F(a h) = hIf'(a + Oh) + f'(a-Oh)]
(fat h)-2f (a) + f(a—h) = hif'(at oh) + f'(a- oh)
Solution.
(a) Define F: (0, 1] > R by
F(x) =f (a+ hx) -f(a— hx)
‘Asx varies over 0 to 1, a + hx varies from a to a+ h and a ~ hx varies from a to a ~h. Thus,
Fis continuous on [0, 1] and differentiable in (0, 1).
Also, F(0) = 0, F(1) =f(a+h)-f(a-A)
and F(x) = hf'(at hx) + hf"(a- hx)
Hence, by the Mean Value Theorem there exists 0 € (0, 1) such that
F(1)- FQ) _
=a (0)
= Slath) —f(a-h) = hif'(a+ 9h) +f'(a— OH]
() Consider the function G: [0, 1] > R defined by
GQ) =flat hx) +f(a- hx)
and proceed as in (a)
6. Show that for 0 tan} b-tan a= (b—a)- z
14
' loo 1
Bow, aees” Ita 14? tae
b-a b-a
—— > Ve 1+a”
4
Taking a= 1,b = 3, we get
4
-1]—]_tan-| i
ar “tan (3) tan “M()< 4
T+
(3)
3 4) x01
= ate ctnt tele,
25° 4 3°64
7. Show that 2 R be defined by
sine 49
L@=) *
1 x=0
Then, fis continuous on [0, 7/2] and differentiable in (0, 1/2).
x cos x~sin x
Now, SQ) = z a
We write *
g(x) =x cos x — sin x
> &' (x) = cos x—x sinx—cos x
=~xsin x <0 in (0, 1/2)
= gis strictly decreasing on [0, n/2]
> ax) S'@)<0 for all x € (0, x/2)
=> fis strictly decreasing on [0, n/2].
= SO) >f(e)>f(m/2) for all x € (0, n/2)
= 1> x2 toe < ®.
x on 2
= erent ford
& prove that > Sin x
for0 0 for all x in (0, 7/2), therefore we need to show that
tanx sinx—x?>Q
for all x € (0, n/2)
Letebe any real number in (0, 7/2),
Ln S)=tanxsinx=%2 forall x € (0,
Then fis continuous on [0, ¢] and differentiable on (0, c)
Now,
f'(&) = sec* x sin x + tan x cos x—2x
= sin x (sec? x + 1) 2x
The form of f(x) is such that we cannot decide about its sign.
Let g@)=s'a)
= sinx (sec? x + 1) ~2x, x € [0, ¢]
Then, 8 (x) = cos x (sec? x + 1) + sin x. 2sec? x tan x -2
= sec x + cos x-2 + 2sin? x sec} x
= (fsee x — [cos x)? + 2sin? x sec} x
> g (x)>0in (0, c)
> gis strictly increasing in [0, cl].
ad &(x)>g(0)=0 for all x € (0, c]
ad f')>0 for all x € (0, c]
> fis strictly increasing in [0, c].
> f@)>fO)=0 for all x in [0, c]
> S@)>0 for all x in (0, 2/2)
> tan x x
> — for all x in (0, 7/2).
x sin x.
* Show that for a> 6> 0 and m N,n>2,
a" pin < (q— py" <(a+ by < allt + lH
Soluei . .
Mon. we write, 1 p, then the given inequality may be written as
@—W<(a— by <(at by 1
> O18) =p (ep —pr
-—P PL
~ (x=phPe xhP
Under the conditions x > | and 0 0.
Hence, 9 is a strictly increasing function in [1, + 2).
> b()>O(1) forx>1.
> o(a)>0
> (x-1P>x?-1
we get
yum gin pin
Putting x = and p
> (a-
For the part (a +b)!" c, then a )
fom (
S'G8) = ne ne) Clem
=2n(x-c)>0
Ifx 0 is a constant
(c) A(x) = cos x on [0, 107]
(og) =412= 12 [on 2, 3]
le’s Theorem is applicable. If applicable,
the derivative vanishes. If not applicable,
/x, x#0
1
wror={ 0, rao Onl).
»
. Prove that between any two real roots of e* sin x = 1 there is at least one real root of
ecosx+1=0.
=
1. Using f(x) = (x — 4)In x, prove that the equation xIn x = 4 — x is satisfied by at least one
value of x in (1, 4).
. Use the Mean Value Theorem to prove that
(@)|sinax—sin bx|<|a—b||x|,x#0 (b) | cos ax —cos bx | < la—b||x|,x#0
© tan“ x-tan4y|s|x-yl,nyeR (det (x-a)0 Oko- Ta
x-a
(© f)= —,a4#b,x4b
x-b
Also indicate the intervals on which the function is increasing and on which the function is
decreasing,
=
’ Let fbe continuous on [a, b], differentiable on (a, b) and let f(a) = 0 and /"(x) # 0 for all
*€ (a,b). Show that x = ais the only zero of f(x) on [a, ).
Suppose fis twice differentiable on R and that there are three distinet points x
Reith £61) = f(x) =f Gy). Prove that there is al least one point ¢ for which /"(c) = 0.
xy and x3 in
aDIFFERENTIATION-1
f ; i
9. Suppose fis continuous on [a, b], differentiable on (a, 4) and that L(x 0
. ©
constant. Prove that /(x) = cx +d on [a,b] for some Constay is ally,
1 ant
10. Prove that, if fis differentiable on (a, b) and "is bounded on (a, h), :
continuous on (a, b).
IL. Let: R > R satisfy the condition | f(x) -0) | < @-y)
aconstant function,
(a, b), where c
the
en fis Pilon
Y
P for allx, y ¢ R. Show that
i
12. Prove that the function f(x) = x + x? sin (2/x), x # 0, /(0) = 0 increases at x =
0, but q
increase on any interval J = (~¢, ¢) for any ¢ > 0. CS ng
13. Suppose fis continuous on [a, b] and differentiable on (a, 6). Then show that fo
& defined by g (x) = f(x) + Ax, x € [a, b], where A is chosen under the con
g (b), there exists c € (a, b) such that g’(c) = 0.
Hence show that f(b) =f(a) = (©) (b ~ a). (This result provides an alternative pra
the Mean Value Theorem). a
14, Suppose /: [0, 2] -> R is differentiable. /(0) = 0,f(1)=2 and f(2) =3. Show that
(a) there exists c; € (0, 1) such that f'(c,) = 2;
(6) there exists cp € (1, 2) such that f’(c:) = 1;
(c) there exists c; ¢ (0, 2) such that f"(c3) = 3/2;
the function
dition » (a
15. Use the Mean Value Theorem to prove that
(Ae 1,
a
. (Bernoulli's Inequality) For p > 1, show that
(1+xP21+px forallx>-t.
. Show that e* > 1 +x for allx € R,x #0.
3
—— = ASSORTED PROBLEMS eee
1. Let fbe a continuously differentiable function on R. Suppose that L= lim (f(x) +f")
ae
exists. If 0 < L R is
about the function /?
-mpowine is
ven by f(x) = el*14 + [x2 — 1], Which of the followin’
(A) It is not differentiable exactly at three points of R.
(B) It is not differentiable atx = 0.CHAPTER 3
| # his agifterentiable at x = 2
oO fferentiable atx = 1 and x =—1,
pyitisnet a INET Dee, 20134
ition AS) 40 HLF AP dierent atx = 9
© ay oro values OF i Ay Aas ay (B) for any values of a, asa, 4,
oO only ifn = 0, (D) only if both 4)=0anday 0,
pe . INET June, 20134
a saps /:R> Ris a fetion that satisfies 70) -/()| <[x-y,B>0, wien of the
following, is correct?
oa a a (B) If B > 0, then fis uniformly continuous,
(up> 1, then fis a constant function. (D) fmust be a polynomial. (NET June, 2013)
5 Letfbea twice differentiable function on R. Given that f"(x) > 0 for all x € R,
(A)f(@) = 0 has exactly two solutions on R.
(B) f(x) = 0 has a positive solution if /(0) = 0 and if") =0.
(Cf) = 0 has no positive solution if (0) = 0 and f"(0) > 0.
(0) /() = 0 has no positive solution if f(0) = 0 and f"(0) <0. (NET Dec, 2012]
3
2
6 Leif) =sin x —x + a and g(x) = cos x— 1+ a for x € R. Which of the following
statements are correct?
(A)f(@)20 for all x>0 (B) gis an increasing function on (0, «0)
(C)gisa decreasing function on [0,°) _ (D) f" is a decreasing function on [0, «).
[NET Dec, 2012]
1. Consider the function f(x) = cos (|x —5 |) + sin (|x —30[) +|x +108 -(x|] +4)?
At which of the following points is fnot differentiable?
(A) (B) x=30
(x=-10 (D) x=0. [NET June, 2012]
§ Consider the function Ax) = | cos x | +| sin (2 -x)].
At which of the following points is fnot differentiable?
(A)
ns n5 nec} (B) {naineZ}
Otmnsr:ne zy (D) {= ine 2 [NET June, 2011]
da pe as
eT ER RY) —/) |< k [Oe—y)] aM) forall x,y € Band for some a>
Sor i
. ™ek> 0. Which of the following is/are true?
“Ye F is continuous, (B) every fe F uniformly continuous.
De . ae i
| “SY dileentable function fis in F. (D) every fe F is ilterentable.
(A)
[NET June, 2011]
ngDIFFERENTIATION-1
12
10. Leta> 1 and fg, h: [-a, a] -> R be twice differentiable functions such that for
Some,
with 0 R which is twice differentiable and sati
x €[0, 1], we must have
(A) f"(@) 20 for all x € [0, 1].
(B) f'(%) 2x forall x € (0, 1).
(C)f G2) —x2 < f(x) —x for all xy, x» € (0, 1] with x) > x).
) f (42) —x2 2 F(%1) x; for all x, xy € (0, 1] with ay > x,.
VAM ans
(AM 2006,
sfies f"(x)> 1 foray
(JAM 2007)
13. Let be a differentiable function on an interval (a, b). Assume that "is bounded on a, 8)
Prove that fis uniformly continuous on (a, 6). (JAM 2009)
14. Suppose f: [0, 1] — Ris differentiable and f"(x) < 1 at every x € (0,1). Iff(0)=0 andf
(1)= 1, show that f(x) =x forall x ¢ [0, 1]. [JAM 2011)
15. Suppose fis a real valued function defined on an open interval I and differentiable a every
xe L If (a, 5] cl and/"(a) <0 R be a twice differentiable function such that” is continuous at every point
in (a, b). Prove that
tim LE+D=2. fe) +f(x-h)
x30 h
for every x € (a, 5).
=f"e) (ua 201l]
17. Let f: [a, 5] - R be a continuous function such that fis differentiable in (a, c) and (¢,4)
a R be a differentiable function such that f(x?) = 1 x3 for all> 0 and (1)
Then f(4) equals
41 47 16
— B) — 16 8
As ® 4 O-% (D)~} (aM 2014)
a . . 1
ni. Let fs RAR be a twice differentiable function such that f (3) =0 for all n & N. Show
that f"(0) = 0='"(0). JAM 2014)
22. Letf: (0, 1) > R be a differentiable function such that | f’(x)| <5 for all x € (0, 1). Show
1
that the sequence {r (ah converges in R. JAM 2014]
2B Let: RR be a differentiable function with /(0) = 0. If for all x ¢ R, 1 "(x) <2, then
which of the following statement is true on (0, 0)?
(A) fis unbounded (B) fis increasing and bounded
(C) fhas at least one zero (D) fis periodic. JAM 2015]
44. For n> 2, let f,: R > R be given by f, (x) =x" sin x. Then at x = 0, f, has a
(A) local maximum if n is even (B) local maximum if n is odd.
(C) local minimum if n is even (D) local minimum if mis odd. (JAM 2015]
8. Form,n€ N,
x” sin (4) x#0
In, n &) = x
0, x=0
( .
. differentiable for each pair m, n with m > n.
) differentiable for each pair m,n with m n.
[JAM 2015]
hat / has finitely many
‘Then for any y € R, the
[JAM 2015]
entice for each pair m, n with m R be a twice Continuously differentiable function with LO)=F0)
Then
LO) 4
(A) f" is the zero function
(B) £"(0) is zero
(OF) = 0 for some x € (0, 1)
(D) f" never vanishes. INET Deg M5
28. Letf: R>Rbea differentiable function such that sup |f*(x)| < ©.Then i
xER
(A) Smaps a bounded Sequence to a bounded sequence
(B) fmaps a Cauchy sequence to a Cauchy sequence
(C) fmaps a convergent sequence to a convergent sequence
(D) fis uniformly continuous, INET Dec, 29151
1
7 for all x © [-
1
29. Let f(x) = —T 1+fe-q]
1+
1, 1]. Then which one of the following is TRUE
(A) Maximum value of f(x) is 3 (B) Minimum value of f(x) is :
(C) Maximum of f(x) occurs at ; (D) Minimum of f(x) occurs at x = 1,
(VAM 2016)
30. Let the function f: R > R be defined as
x, if xe(0,2)N@
SQ)= .
2x-1, if xe (0,2) (R/Q)
which one of the following is not true
(A) fis continuous at x = 1 (B) fis differentiable at x= 1
(C) fis not differentiable at x = 1 (D) fis differentiable only at x= 1.
[DU Ent. 2016)
31. Let $: ROR be a differentiable function such that ' is strictly increasing with (1)
= 0. Let a and B denote the minimum and maximum values of ¢ (x) on the interval [2, 3].
respectively. Then which one of the following is TRUE ?
(A) B= 6G) (B) a = 9(2-5)
(C) B= (25) (D) a = $(3) [JAM 2017]
l+x if x<0
32. f= {4 oGeag if x20 a
satisfies the assumptions of Rolle’s theorem in the interval [-1, 1], then the ordere
(p, 9) is
(A) 2 -) (B) (2, -D
M2017]
©) ee ue18
sin 1X CHAPTER 3
get) sing? * © (+) and tet x © (0, ) be such that
(f(00))? + (2? = 1) Sin? x9) = oenrne (0) = 0. Then
[YAM 2017)
3c (1
x sin| —
44, Fors € R, let f@) = (=} x#0
y >» x=0
Then which one of the following is FALSE?
£0) 4
A) lim 2
o) a0 *
Doe. .
(¢) L& has infinitely many maxima and minima on the interval (0,1)
7 ,
(D) LO) is continuous at x = 0 but di i
£2) sat x = 0 but differentiable at x = 0 [JAM 2018}
x
35. Let f: R > R, be defined by
sin x
f@=} x’
1 >
Then,
(A) fis not continuous
(C) fis differentiable
(B) fis continuous but not differentiable
(D) fis not bounded [NET June, 2017|
1
36, Letf: R/{0} > R be defined by (x) = +3
On which of the following interval (s) is fone-one ?
(A) (, -1) (8) @, 1) on
© ,2) (D) (0, ) [JAM 2018]
3. Let: R, be such that f” is continuous on R and f(0)
Then, lim (( 2) is...
i400 x
=1,f"(0)=0and f" (0)=-1-
i 018}
. (correct up to three decimal places) [JAM 2 ]ES
eae
< Differentiation—-]]
In this chapter, we further continue our study carried out in the previous chapter
differentiable functions. We have come across several applications of the Mean Valye Theore
in that chapter. Here we introduce two generalizations of the Mean Value Theorem oy.”
Taylor s Theorem and the other is Cauchy's Mean Value Theorem. While the Mean Valye Theorem
relates the values of a function and its first derivative, Taylor 5 Theorem provides a relationshin
between the values of a function and its higher order derivatives. On the other hand, Can
whys
Mean Value Theorem deals with two functions together along with their first order derivatives
about th,
TRTAYLOR’s THEOREM
Taylor’s theorem provides us a way to approximate a given function by a polynomial
Approximating a function by a polynomial is a very useful technique in real analysis. Itis because
the polynomials are one of the simplest kind of functions and operations such as differentiation,
integration etc. are easy to compute for these functions. Taylor’s Theorem gives us the tool to
look for the suitable polynomial for a given function and also describes the error involved in
the approximation.
Definition 4.1.1. Suppose fand its irst 2 derivatives f’, #”, ..., f exist at a point a
Then the »/” Taylor's Polynomial for f about a (or at a) is de ined by the expression |
L@ LO, |
n!
(c-a) +... (el
T, @)=S(@) +f'@ (xa) +
2!
Example 4.1.2. Find the 3rd Taylor's Polynomial for the function f (x) = sin x about 0.
Solution. Let f(x) = sin x. Then
f@)=sinx => f(0)=sin0=0;
S'@)=cosx => f'(0)=cos 0
S'@)=-sinx = f"(0)=-sin0=0;
S")=-cosx => (0) =-cos 0
£0) 3, £70)
T3 (x) =S(0) + f"(0) x + 3 3
1264a CHAPTER 4
Ox? =)
=O+1xe+ 42
~ 2
13
=x 5
6°
gxample 4.1.3. Find the 4" Taylor s polynomial of the function
Solution. Here f(x) = 3 + 5x 4x3 + x4. Then (1) = 5 and
SO)=10K- 122 443 = pyar.
I") = 10-24r 4122 ry
SX) = -24 + 24x => f"(1)=0;
SOK) = 24 => SOL) =24
(Xx) = 3 + Sx? 4x3 +4 about |,
Hence,
THA)= FO) +/'Me-1)+ FO eps LO gay
(4)
+£0 (x- 14
0 24
=5426-D- 5 @aseed @-1Ip+ @-14
=5+2e-1)-(@-1P +=,
Example 4.1.4. Find the n™ Taylor's polynomial for the function f (2) = e* about 0.
Solution. For each n € N, the n derivative of fis
S) (x) = ef and £0) = 1.
Hence
T,@)-fO+POx+L 2° v++ £0 o e
x”
xe
Bltet gy tt
How far does a function f differ from its Taylor's Polynomial? The answer lies in the Taylor's
Theorem provided below.
Theorem 4.1.5, (Taylor’s Theorem with Lagrange’s form of remainder) Let I = [a, 5], f:
> [Link] € 1. For n € N, let fhave derivatives upto 7 order on I such that
© is continuous on [a, b] and
(f° is derivable in (a, b).
Then for any x € I, there exists a point c between x and xp such that
"(y)
10) = Sa) f°) @—20)+ FED (aay.
-(n) (ml) (0
‘ ff! Me) LO) (e-x"!
n!
G20)" * GeDIFFERENTIATION-II
a a oo :
Proof, Let J denote the closed interval with end points x and x9. Define a function 8
F:J->Rby
“ LO (gL
F=f) | LOF OOD 475) (8-0 tt O(N NG in
Aisa constant to be suitably chosen.
Now, Jc [a, 5]; hence fhas derivatives upto n'* order in J.
Also, is continuous on {a, 6] and derivable on (a, 5); hence, f JF, fg
continuous on [a,b]. Therefore F is continuous on J. Similarly, F is derivable on J, except bins
at the end points of J (since x, x9 may also take values a and b, F may not be derivable atx ny
x). Further we choose the constant A so that F(x) = F(). Then F satisfies all the condition
Rolle’s Theorem. Hence there exists c between x and xp such that F’(c) = 0.
where
HD) (¢ 1)"
p= LO Oe ) -A(n+1)(x-0"
Now,
Hence, F()=0 > A Le
(n+l)!
Again, Foss) = FG)
Sl6)+ f'GNe=%)+ 222 (ea) +.
= fe@)- 7 a =0
L008) ayy Abe"
= Fl) =f(00) +f) (ex) ++ OO (ex)?
£0 (40) (OC ga
tot rie (=x) + oo (r-n
Remark
If we denote the 1!" Taylor’ , .
may be written as vi® Taylor's Polynomial of fabout x by T,(x), then Taylor's Theorem’ oom"
S()=T, (2) +R, (x)
Le)
(n+l)!
‘ The expression R,, (x) is called the Lagrange’.
Several other forms of remainder for ‘ature, We shi
fone more such form, known a “
where R, (x)=
(exp)
yr
form ofrems™
for some point c between x and ry.
ise
‘STI fn the Il all
\ylor’s Theorem are known in the lit
_—_——CHAPTER 4
429
ing Taylor's Theorem prove that
Je 4.1.6. Using Tay
Examp ? A
+ FE a ees x > 0,
I+x 23 rary e 7
x> 0. We apply Taylor’s Theorem to the function f(x) = e* about 0, with 7 = 2 to get.
e = Ty (x) + Ry (x)
xe
Tr (x)= 14tx4+ 7
solution. Let
where
-L0
and Ry (x)= ai x
e
acy x3, for some c between 0 and x.
Since, 0 TL@+ 3 <@)+ a e tet tay se x
Example 4.1.7. For any k € N and for all x > 0, show that
ack,
2k+1
Solution. Let x > 0. We apply Taylor’s Theorem to the function f(x) = In(1 +.x) about 0 to get
In(1+x)=T, (0) +R, @)
rete Ft peeing tay S@O= a LO=1
1
"(x)= -———, 0) ==1
fa=-T ae f'O)
ml
700 y= DD! poy) 211)
Hence, (+x)
LO) Os £2,
nt
T, (x)= S(O) +f" (Ox +DIFFERENTIATION-I in
ie
cix- batt
=O+1x¥ 3 . |
(0) yy EDTA I .
™ RO “Gan ~ dre" ep!
en
= aaa aed x"! for some c between 0 and x,
(ite) (at
Now, foranyx>0, R,@)>0, — if'mis even
and R,@)<0, — ifmis odd
Hence, S (8) = Tae @) + Rox &) > Te @)
and S(&) = Tres @) + Roser ©) < Tre)
Therefore, from (i), we get
2 xk 1
yo +..- 2 Rby ih end points x and x9 by J and define a funt
6 O=Se)- [7 0+ £O@-+ L0G
(n)
et LO (pnt eae
nl
‘here A is a constant to be suitably chosen
Then, as in Theorem 4,1
the end points of J Further, fark Asean co en nino
all the conditions of Rolle’ Thee the constant A so that G (x) = G (a). Then
's Theore erel
such that rem. Therefore, there exists som
5., G is continuou:
and 80
e ¢ lying betwee!CHAPTER 4
431
yon
or G(n= fo (x- "A
‘Therefore, a
G()=0 > A= fa (e-oy"
nl
Again G (x) = G(x)
Dae [: (30) + 1°09) (= 59) +22) (6-5)?
)
+O (en) FAG) 0
. I0)= $0) +76) (0-9) ++ LO) 65 y
on (mst)
tet L&) (=x) + oo (x= c)" (x9)
Remark
1£0 € I, we may take xq =0. Then the number c obtained in Taylor’s Theorem lies between 0 and x,
so that we may write c= 0 x for some @ with 0 <6 <1. The expression of f(x), then, can be written as
” )
Fe-fO +f Ox tO) 244° O ee ey 0g
2! nl
(n+l)
where RG) = eer re (Lagrange’s form of Remainder)
n+l)!
(n+)
or Ry @)= eek) (1-6)"x""!_ (Cauchy’s form of Remainder)
n
The particular case of Taylor’s Theorem with x9 = 0, fhaving the above expression (i), is also known
4s Maclaurin’s Theorem.
<=“ ILLUSTRATIVE EXAMPLES. ———
1 Ife (0, 1] and n & N, show that
n+l
2 3 n
marn-(Fet—rcor se] <
ae ‘his to approximate In 1.5 with an error less than 0.01.
Tavory a write (x) = In (1 + x), Then f possesses derivatives of every order. Hence, by
corem, we get about x = 0
S(x) = T,, (x) +R, (x)
ro)
“(O) 4 ae) 7
= £0) +f") x4 ao eh eR OD
MereDIFFERENTIATION-I
(nil) (e)
where RO “Gap!
roe oO) = CA Hor DYE, Hen
2 x?
In(l tay= 04x + 3
= [ings o-[ wt CD
—, <
n+l(l+q"™
Now, for In (1.5), we have x = .5
For m= 3, we have
ny
vy"! for some ¢ wilh 0 6 l
ntl
atl pad
wat Ld ~ o1s625> 01
n+l 4 2.4 64
Then, for n= 4, we have
wt Ft LN 0605.01
n+l 5 25.5 160° :
Hence, the error is less than 0.01 if m= 4.
Now, for n = 4, we have
= 5 —.125 + .041666 — .015625
= 541666 — 140625
= 401041
1. Show that
2 24
ne?
| Scosxs) + for ally
< S Fall x €
2 24
2. Show that sin x lies between
3 ‘ s
x :
x-—andx-* 4%)
6 6 120
R.yr
FT 33 CHAPTER 4
Taylor's theorem to prove that
nue
x . ‘i
XS cet
+5 HO,
1+
prove that
sin a mu
irfla) =e show that the remainder term in Taylor's Theorem
sp for each fixed xp and x,
9
x
— forall x,
1 II x,
”
converges to (as m tends to
_ gfe) =sinx, show that the remainder term in Taylor's Theorem converges to 0
‘ as n tends
10 ®.
1, Calculate ¢ correct upto four decimal places,
8, Find the Taylor's polynomial T,,, (x) and T344, about 0 for f(x) = cos x.
4, Find the sixth Taylor’s Polynomial T¢(x) for the function f(x) = J about 1. Also fnd the
Lagrange’s form of remainder Re (x).
10, Find the sixth Taylor’s polynomial T(x) for the function \f (x) = Inx about 1. Also write the
expression for the Lagrange’s form of remainder Rg(x).
AUCHY'S MEAN VALUE THEOREM AND L’HOSPITAL’S RULE
Now, we come to the other generalization of the Mean Value Theorem — Cauchy's
Mean Value Theorem. As an application to this, we will prove a very useful result of calculus,
known as L’Hospital’s rule.
Theorem 4.2.1, (Cauchy’s Mean Value Theorem). Iffandg are continuous on §,, b] and
differentiable on (a, b), then there exists c € (a, b) such that
(6) -f(a)] g' (c)=[g ()-8 MISO.
Proof. We define h : [a,b] >R
h(D=[F()-SlgO-[g)-s@OlSO i)
Then, h is continuous on [a, b] and differentiable on (a, 6). Therefore, by the Mean Value
Theorem, there exists ¢ € (a, b) such that
h(b)—h (a)= ht (c) (b-a)
Now, from (i)
h(b)= f(b) g(a) ~g (6) f(a) = h(a).
Therefore h(e)=0
Hence, we have
VO)-fla) y ()={g(6)- KAI {Krom (OLDIFFERENTIATION.
Remark
HWwe take g (yy Vforalle fa, by
MY Hay (b
Which is the conclusion
Uhus, 1
Hin Cauchy's Mean Value Theorem, then we Bet
are)
ofthe Mean Value Theorem,
c's Mean Value Theorem is a part
ally, suppose x IY = east Cquation of a ¢,
ave dierentiable on (a, then for any fe (a, h ely
nb), i *0, the Slope of th
© Curve
gO
POINL (A), (y © Cauchy's Mean Va
r (HO. 2 (0) ism rw The Cauchy s Mean Value Theorem thus Says that Under
there is some ¢ © (a, 5
al Lo the slope of the secant
curve,
Aran
Goometr
andy
4 is parametric
), Where /"(
atthe
the given conditions,
VO 8 CD) is equi
) for which the slope o
#(h)) of the
f the tangent at the poiny
line through the endpoints (/ (@),g (@)) and (by,
Example 4.2.2. Show that
sina =
inf
I cot), for some
08 B cos
0 € (&, B), where O<
Solution, Let f(xy
a
[eos cos 0 # 050 0 and g
Le B(x)
(*) > 0, when x > ¢:
. f(x)
or lim AB)
+ Where both /(x) > co and g
ve gy
(%) > ©, when x ¢,
‘Types of such quotient : are called indeterminate forms; while the first type is symbolized
as © form, the second one is symbolized as = form. Other indeterminate forms are Ose,
Isa
However, limits of all these indeterminates can be evaluated by
sing algebraic manipulations,
0
° asic Ie is — or —.
logarithms or exponentials rules to convert them to the two basic form: E7
CHAPTER 4
438
Pe provide proof of L'Hospital’s rule for the basic indeterminate form 2 . St may
pelos ; oo 0
tioned pere that there are several other versions of L'— Hospital's rule in the literature.
, ;
pe ne 0
m 423. (LHospital’s Rule (basic) for form). Let fand g be continuous on [a,b]
orem . «
™ giferemtiable O° (a, 6). Let e € [a, b] such that f(c) = 0 = g (c), Further let g(x) #0, g(x)
oll ‘re (a, DMC}. We have
20! .
im £9 21 then tim LOD
if soe g(x) xe g(x)
rons, Weappeal othe sequential criterion of limits. Let (x,) be a sequence in (a, b)\{c} which
mets ‘oc, By the Cauchy’s Mean Value Theorem, there exists c, between x, and c for each
such that
fe) -SO1 8 Cn) = Ig On) - 8 OF) sali)
Now,@(2) and g’ (x) are nonzero for all x ¢ (a, 6)\(c}. Therefore g (x,) # 0 and
go) #0 forall n. Also, f(c) = 0=g (c). Hence we get,
Ln) _fG)-fO _ f'n).
B%,) 8%)-8lc) 8'(C,)
Since, ¢, lies between x, and ¢ for each n, and x, — c, we have c,, > c. Hence, by the Sequential
F lim f2.-1)
sve g(a)
[by (0) Mid)
Criterion for limit
7 (Ze)
8'(Cy)
> lim (48 [from (ii)
> lim
rove g(x
[Remark
Thesatement of Theorem 4.2.3 for 9 form is provided in a manner such that Cauchy's mean Value
Theorem can be directly used in the proof. Other simpler statements also exist for L’Hospital’s rule for
= fi 20
a om as Well as for = form, We provide them below here.
Theo
rem 4.2.4, (L’Hospital’s Rule)
0
@{o
9 rm | : 1ffand g are two functions such thatOIFFERENTIATION«I|
(9 Tim 700) = tim g (9 ~ 0; "
(OFC), g(r) exist
and g(x), 20) #0 forall xe (c
some 8 > 0 and
det 6) except POscibly ay
")
(x)
(ii) lim
reg
ssthen, tim LU = jing £0).
Xe Q(X) xe a(x)
()) (= form) IFF, gare two functions such that
(i) tim S(x)= lim gat © (or ~ 00);
(i) f'(), g'() exist and g
(), 8'@) #0 forall x € (¢~
some §> 0, and
8,c+8) eXCept possibly atc. fo
Git) tim LO) exists: then, tim LO) — tim £@)
xe g’(x) xe g(x) xc g’(x)
Remark
The above result also remains valid if limit is teplaced by rij
‘That is, we can replace c by c* or c~. In that case,
(c,€ + 8) or (¢~8, c) respectively.
ight hand limit or left hand limi
(C8, € + 8) of (i) should be suitably epiaceq by
P. . fanx—x
Example 4.2.5, Evaluate lim
x30 x? tanx
Sok jim {92% _ 5, (tanx—x x }
olution. lim ——— = on eae
roo tanx x90 tan x
tanx-x (0 i es
= jim 222% 9) |. him =1
im, 3S (3) ( x0 tanx )
aT sec? x-1 3)
so a (0
2sec? x tan x
= lim
roo 6x
1. a (=)
= ~ lim sec? x] =
x0 x
1 1
= sles
30043e CHAPTER 4
ILLUSTRATIVE EXAMPLES
sin 2x + asin x
1, For what value of a does ear nae tend to a finite limit as x > 0?
__ sin 2x+asinx 0
sation. I? 0
= jim 20082x+acosx (24a
x40 3x2 0
qoapply L’ Hospital’s Rule, we assume 2 + a=0
shen the above limit = jim 20082x=2e0sx (0
’ x90 3x2 0
= fim, TAN 2x4 2sin-x (0
x0 6x 0
= Tin oem ae Aas
x90 6
=-1
Hence, the required value of a is -2.
2, Evaluate lim [ c 1]
xo0le*-1 x
Solution. The given limit is of 0 ~ « form.
pet
cry pe ee ree (2)
xo0(eF-1 x) 190 x(e*-1) 0
Sew (@)
roe —1+xe" \0
= li
x
. -e
= lim ———;=-5
x90 et +e +30! 2]
——— (EXERCISES 4.2
¢ following functions and
1. Check the applicability of Cauchy's Mean Value Theorem for th
Verify the hypothesis in case of applicability
(9) f0)=x-22, g (x) = 3x2 4x? in [0.1]
FC) =e, g(x) =e in 10, 1]
(©F(4)= sin x, g (x) = cos x in [-7/2, 0.DIFFERENTIATION-II
, P "
2. Find c as prescribed by Cauchy’s Mean value Theorem for :
(a) (x)=, g (x)=? in [1,2]
(Df) =x (= 1)r—2), g (0) =x (e= 2)or~ 3) in [0, A].
(Of) =e, g (x)= e* in [0, 1]
3. Evaluate the following limits
x :
(@) tim 22+) 8) tim 2!
2 © ine, OF
(©) Kim (sin x)"* (@ lim, (+x)!*
wi
(@) lim (tan x)" (D lim, (cot x)*™*
x0"
oe
2
4, Find the values of a, b, c so that
. ae* cos x +c ris
Jim 42 =P 008 ** 06 may be equal to 2.
2390 xsinx
5. Find the values of a, b, cso that
a+bcosx+csin x
lim 7
x30 x
FEMI SOME APPLICATIONS OF TAYLOR'S THEOREM
In this section we discuss some applications of Taylor's ‘Theorem with regard to
convex functions and relative maxima.
exists and equals ; .
Convex functions
Convexity is a very important mathematical notion which has a number of applications,
particularly in the Theory of Optimization. Graphically speaking, a function f: 1 Ris convex
ifthe graph off between any two ofits points lies below the line joining the points. ‘An immediate
example of such a function is f(x) = ||.
Definition 4.3.1. Let | cR be an interval. A function f: I + R is said to be convex on lit
for any with 0 <¢< 1 and any points x), x2 € I, we have
F((1 1); +e) $ (1-9 Fs) + F)-
Here you may note that for x; <2, O R have second derivative on l Then
a
Convex function on | if and only if f"(x) 2 0 for all x € L
Too}
I. Let be a convex function on I. First we shall prove that
+m-2f@+fe-"
f"a)= lim Lei
nyo
points. However,
except possibly at the end
perties exhibited