Chapter 2
POLYNOMIAL APPROXIMATIONS
Let f be a function having n finite derivatives. We shall give 1i this chapter an estimation of the error in the following polynomial approximation:
n n
a x n
a f a
x a f a x a f a f x
f ( )
! ) ( ...
) ( ! 2
) ( ) ( ! 1
) ( ) ( ) (
) ( 2
" '
. a
1. PRELIMINARIES
Let I IR be an interval, a I , and f :I IR be a function. We remember that if there exists:
) ( : ) ( : ) ( ) ( lim )
( ) (
lim '
0 dx a
df a f h
a f h a f a
x a f x f
h h a x a
x
,we call this limit (finite or infinite) the derivative of the function f in a. In the following paragraphs we shall use some known theorems.
Proposition 1: Any function having a finite derivative in a point iscontinuous in that point.
In fact, this Proposition asserts that, in a neighbourhood of a point a, the function f that has a finite derivative f
a can beapproximated by a first-degree polynomial:
) a x )( a ( f ) a ( f ) x (
f ' .
One of our purposes is to generalize this approximation. We shall remember some derivatives for elementary functions:
uv ' uv1vu' uvv' lnu, whereu, v have derivative, u>0
chx2 e e 2
e e shx
x x ' x x '
Rolle’s Theorem. Let f :
a,b IR be a Rolle function (i.e. fis continuous on [a,b] and has finite derivative on (a,b)), such that
f(a) = f(b). Then there exists a î
a,b such that f'( )0.
Fermat’s Theorem. Let f :I IRIRbe a function. If a is an interior point of interval I (i.e. å>0 such that (a-å, a+å) I)
which is a local point of extreme for f and f has a derivative in a, then f'(a)0(i.e. a is a stationary point).
Primitive’s Theorem: If f :
a,b IRis a continuous function then:
x
a
dt t f x F b
a
F: , IR, ( ) ( )
is a function for which F’ = f.
Mean Value’s Theorem: If f :
a,b IR is a continuous function, there exists a point a î
a,b such that). a b ( ) ( f dx ) x ( f
b
a
2. DIFFERENTIABILITY
Definition 1. A function f :I IR is said to be differentiable
in a point a, if there exists a constant AIR and a function IR
I
:
continuous in a, with (a)0, such that
) x ( ) a x ( ) a x ( A ) a ( f ) x (
f
for all xI; the linear function: da f :IRIR,daf(h) Ah
is named the differential of the function f in the point a.
Proof. The necessity. From the hypothesis we know that there is AIR such that
I x ), x ( ) a x ( ) a x ( A ) a ( f ) x (
f (1) where (a)0and ù is a continuous function in a. From relation (1) we can write that:
a x
a f x f a
f
a
x
) ( ) ( lim ) (
' lim
( )
,) 1
( xa A x A
since ù is a continuous mapping; therefore f has a finite derivative in a: A=f’’(a).
The sufficiency. Suppose now that f has a finite derivative in
a, i.e. the limit:
) ( ) ( ) (
lim f' a
a x
a f x f
a
x
,
exists and it is finite. To prove that f is differentiable in a, let IR
I
:
defined by
), a ( f a
x
) a ( f ) x ( f ) x
( '
if xa, and ù(a) = 0; (2) then:
) ( 0 ) ( ) ( ) ( lim ) (
lim f' a a
a x
a f x f x
a x a
x
;
hence ù is a continuous function in a; moreover, from (2) we have: f(x) f(a)(xa)
(x) f(a)
,hence the function f is differentiable in a and A = f’(a).
Remarks. 1. The differential in a, of the identity 1I idI is denoted by dax, and dax(h)h; obviously daxdbx for all b; therefore, the differential of the independent variable x (i.e. idI) is denoted by dx; with this notation we can write the differential of f in a like this: da f f'(a)dx.
2. If f has finite derivative in every point xI, we say that f is
differentiable on I and
dx
dx df dx f
(where
dx df
f' is the derivative of f in the current point) is named
the total differential of function f.
3. If f and g are two differentiable functions, then
d
f g
df dg(the linearity of the operator d)IR , ,
(if ( ) 0, ) 1;
' '
2
' '
I x x
g df
g dg f g g
f d
df g dg f g f d
d(f u) f'du, for all compoundable functions f and u. 4. We remember that, if f :I IR is differentiable, and f’ is differentiable in a, then:
) a ( f : ) a ( dx
f d ) a ( dx
df "
2 2 '
is named the second order derivative or the second derivative of f; the monomial function of second order:
2 2 " 2
) ( ) ( ,
: d f h f a h
f
da IRIR a
is called the second order differential in a or the second differential of f in a; since
2 2
h ) h (
dx
if f has second derivative on I, then: d2f f"dx2,
where 2 " 2
a f f (a)dx
d and 2 " 2
a f(h) f (a)h
d , is named the
second differential of f (on I).
5. Analogously, if there exists f n (a), finite, the n-monomial function:
, ) ( )
( ) ( )
( ,
: n (n) n n n
a n
a f d f h f a dx h f a h
d IRIR
Example 1. Let f :IRIR, f(x)xexcos2x. Then:
e cos2x x e cos2x 2x e sin2x
dxdf x x x
is the total differential of f; for a = 0 we can write the linear function
f
d0 -the differential of f in 0: d0f (100)dxdx
which is the identity function d0f(h)= h, for all hIR. The second differential can be obtained from the first:
2 2
) 2 cos 4
2 sin 2 2 sin 2
2 sin 2 2 cos 2
cos 2
sin 2 2 cos (
dx x xe
x xe
x e
x xe
x xe
x e
x e
x e
f d
x x
x
x x
x x
x
and for a = 0, we obtain a 2-monomial: d02f 2dx2,i.e. d f(h) 2h , h IR,
2 2
0
which is the second order differential of the function f in the point
a = 0. Analogously:
d03f 9dx3,and ( ) 9 , . 3
3
0 f h h hIR
d
Exercise 2. Find d02009f , where f(x) = x
2chx.
Solution. We have Leibniz’ formula:
n
k
k k n k n n
v n C v
u
0
) ( ) ( )
(
, where .
! )! (
!
k k n
n k
n Cnk
Here, since
2009 2 (2009 ) ( )
0 2009 )
2009 )
2009 ( 2
) ( ) ( )
0 ( 2007
, 0 )
( k k
k k k
chx x
C f
k
x
|x=0 =
= 2( ) 2 ( ) 0 2008 2009,
2 2009 2009 '
2008 2009 "
2007
2009 chx C x chx C x chx x
C
hence
3.TAYLOR`S FORMULA
Definition 1. Let f :I IRIR, having nth finite derivative in a I. The polynomial:
n n
k n
k
k n
a x a f n a
x a f a x a f a f
a x a f k x
T
) )( ( ! 1 ... ) )( ( ! 2
1 ) )( ( ! 1
1 ) (
) )( ( ! 1 : ) (
) ( 2
" '
0 ) (
is named Taylor’s polynomial of n degree, or nth degree Taylor’s polynomialfor the function f in powers of (x-a).
The function
) ( ) ( ) ( ,
:I R x f x T x
Rn IR n n
is called the remainder term of Taylor’s formula, or the nth remainder
(of Taylor’s formula for a function f in the powers of (x-a)).
The remainder Rn(x) shows us the error that appears when we
replace the exact value of f(x) with the value of Taylor’s polynomial
Tn(x).
If a = 0, Tn(x) is named Mac Laurin`s polynomial, and Rn is
called Mac Laurin`s remainder.
We remark that the Taylor’s polynomial can be written as the following:
). a x ( f d ! n
1 ... ) a x ( f d ! 1
1 ) a ( f ) a x ( f d ! k
1 )
x (
T an
n
0 k
a k
a
n
In the sequel we will prove a very important formula- Taylor’s formula- that is a very useful tool in estimating the error that appear by approximation f(x)Tn(x).
Taylor’s Theorem (Taylor’s formula). Let f :I IRIR, be a function where I is an interval, and suppose that f has continuous derivatives of all orders up to (n+1) in a neighbourhood V of a. Then for all xVand *
IN
p there exists a î between a and x (i.e.
min(a,x),max(a,x)
p
For this we introduce an auxiliary function theF:V IR,
( ) ( ) ( ) .
([a,x] or [x,a]); therefore, according to Rolle`s theorem, there exists
(x ) f ( ). p
! n
1
C n 1 p (n 1)
Consequently (1) becomes:
f ( )(x ) (x a) ,
! n p
1 ) x ( T ) x (
f (n 1) n 1 p p
n
and so the Taylor’s formula is proved.
Remarks. 1.The remainder from the Taylor’s formula can be expressed in other forms. For instance, since î is a number between x and a, it follows that î can be written as î = a + è(x-a),
where è(0,1), and if p = 1, then we obtain the Cauchy`s remainder:
); 1 , 0 ( , ) 1 ( ) a x ))( a x ( a ( f ! n
1 ) x (
Rn (n 1) n 1 n
here è depends on n and x.
2. V is a neighbourhood of the point a, hence it results that there is a ä>0, such that [a-ä, a + ä] V; but (n 1)
f is a continuous function on this closed interval, so it results that, there exists M>0: f(n 1)(x) M, x
a ,a
,
and the remainder can be evaluated as:
, ;
)! 1 ( )
( 1
x a for x a
n M x
Rn n
this inequality can be used for many purposes: to investigate the behaviour of the nth order remainder, with n fixed, in the neighbourhood of the point a, or to study the remainder when n
tends to infinity, since: 0 ) (
) (
lim
n
n
n x a
x R
.
Taylor’s integral formula. Suppose that f :I IRIR, is a function that has continuous derivatives up to the (n+1) order. If
I x ,
a , then:
x
a
) 1 n ( n
n (x t) f (t)dt
! n
1 ) x (
)
remainder in the integral form.
Proof. Using the Newton-Leibniz formula and integrating by parts we obtain consecutively:
proved.n n n
n x f a x a
n x
R ( ) ( ( ))(1 )
! 1 )
( ( 1)
which is the remainder of Taylor’s formula in powers of (x-a) having the Cauchy`s form.
4. APPROXIMATIONS OF FUNCTIONS (by
polynomials)
Taylor’s formula is a fundamental tool for approximations. By using this formula we can calculate, almost like with a computer, a large class of functions (values of functions). For instance, if we want to compute f(x) with an error less than å > 0, we know that
, ) x ( R ) x ( T ) x (
f n n
and if Tn(x) , for some nIN, then f(x)Tn(x),
and by that we mean that, when we replace f(x) with Tn(x) the error
is less than å. Moreover, if we know how to calculate
} n ,...., 1 , 0 { k ), a (
f(k) ,
then Tn(x) can be easily estimated with a computer.
Exercise 3. Find a number, nIN such that:
3 1 , 3 1 ,
1000 1 )
(x x
T
e n
x
,
Tnbeing the Mac Laurin`s polynomial for f(x)ex;compute 3 e with
three exact digits (decimals).
Solution. According to Mac Laurin`s formula, for all
3 1 , 3 1
x , there exists a î between x and 0 such that:
, )! 1 ( !
1
0
e n
x k x e
n n
k k
x
and
Exercise 4. Approximate 3 26, with exactly three decimals.
Solution. Let 3
1
) (x x
f . We shall apply the Taylor’s theorem (with Lagrange’s remainder), for 3
3
f can be exactly calculated).
We have:
. 962 . 2 27 80
) 1 ( 1 3 1 3 3 1 3 ) 27 26 )( 27 ( ) 27 ( ) 26 (
26 1 ' 3
3
T f f
Mac Laurin`s Formulas for Most Important
Elementary Functions
The function ( ) x.
e x
f This is a function infinitely differentiable (has finite derivatives of any order) on IR, and
1 ) 0 (
)
(k
f . Therefore Mac Laurin`s formula with Lagrange’s remainder is written as the following:
n
k
n k x
x R n x e
0
), (
! ( ) ( 1)!,
1
n x e x R
n n
where î is between 0 and x. If x
a,a
,a0, then
n
n a e x R
n a
n 0,
)! 1 ( ) (
1
.
This leads to the conclusion that the exponential function can be written as a sum of a series:
,
,
. !0
a a x n x e
n n
x
But a is an arbitrary positive number, so this equality is valid on the entire IR.
Example 5. Let us compute the number e, within 10-3. Solution. By taking x = 1, there is
0,1 such that)! 1 (
3 )!
1 ( ) 1 ( ) 1 (
n n
e R
T
e n n
. It is sufficient to solve the inequality:
3000 )! 1 ( 1000
1 )! 1 (
3
n
n .
This relation is fulfilled for n = 6. In conclusion
2.718,
! 6 1 ! 5 1 ! 4 1 ! 3 1 ! 2 1
2
with an accuracy of 0.001. formula has, in a similar way, the form:
,
Finding Limits with Taylor
’
s Formula
Example 6. Find lim 1
2 3 3 2 6 6ln(1 )
using Taylor’s formula.
Solution. Using the extension of logarithmic function we
therefore
2
), ( !
3 2
1
2 .... ! 4 ! 2 1 2 .... ! 3 2
2 )
(
4 4
4 2 3
x R x
x x x
x x chx
xshx x
f
and
); ( ! 2
... ! 3 ....
! 4 ! 2 1 sin
cos )
(
' 4 4
6 2 4
2 2
2 2
x R x
x x x
x x
x x
x x g
according to Taylor’s formula there exists , '
0,x
such that:
5 ) 5 ( 4
! 5
) ( )
(x f x
R and 5
' ) 5 ( '
4
! 5
) ( )
(x g x
R .
In conclusion, the limit is:
6 1
! 5
) ( 2
1
! 5
) ( !
3 2
1 lim
5 ' ) 5 ( 4
5 ) 5 ( 4
0
x g
x
x f
x l
x
.
Exercise 8. Find
3 4 6 12 12 (ln(1 ) ln )
lim x x2 x3 x4 x5 x x l
x
.
Solution. In order to apply the Taylor’s formula, we denote
x
0,1). (, 5 12
) ) 5 1 ( 6 5 4 3 2 ( 12 12 6
4 3 1 lim
) 1 ln( 12 12 6
4 3 1 lim
) 1 ln( 12 12 6 4 3 lim
5 6 5
4 3 2 2
3 4 5 0
2 3 4 5 0
5 4 3 2 0
0
y y
y y y y y
y y y y
y y
y y y y
y y
y y y y l
y y y y
5. FINDING THE EXTREMA
We remember that the function f :I IR attains its local maximum (minimum) in a point a, if there is a neighbourhood VVa, such that:
I V x a
f x f a
f x
f( ) ( )0,( ( ) ( )0), .
According to Fermat`s theorem, if f admits an extreme in a point a and there exists the derivative in this point (i.e. f’(a)), then
this derivative is zero. By definition, a point aI is called a stationary point of the function f, if there exists a derivative of f in that point and f'(a) = 0.
If a function f is defined on the open interval (á, â), and we need to find all its extreme points, then first of all we should find its stationary points, and then we should search among the points where f has no derivative, if such points exist. Stationary points are to be found by solving the equation:
0 ) (
'
x
f (*)
Of course, not any stationary point of the function f is a local extreme point of f. Condition (*) is necessary for a differentiable function f to have a local extreme in a point x, but not sufficient.
For instance, x=0 is a stationary point for the function ,
, )
( ,
:IRIR 2 1 IN
n x x f
Theorem 2. Let f :I IR be a function. Suppose that there is aIsuch that:
0 ) ( ....
) ( )
( " ( )
'
a f a
f a
f n and f(n1)(a)0,nIN, and also (n1)
f is continuous in a. Then: (a) if n+1 is even then in the point a:
(a1) f has a local maximum if f(n1)(a)0
(a2) f has a local minimum if f(n1)(a)0;
(b) if n+1 is an odd number, then a is not a local extreme for f. Proof. Taylor’s theorem implies the fact that there is a number î between a and x such that:
) ( )!
1 (
) ( ) ( )
( ( 1)
1
n
n
f n
a x a f x
f . (1) Since (n1)
f is a continuous function in a, and f(n1)(a)0, it follows that there is a neighbourhood of the point a, such that:
0 ) ( when , ,
0 )
( ( 1)
) 1
(
a f I
V x x
f n n (2)
or:
0 ) ( when , ,
0 )
( ( 1)
) 1
(
a f I
V x x
f n n . (3)
Case (a): n+1 is even:
(a1) If f(n1)(a)0, then from (1) and (2) we have that:
I V x a
f x
f( ) ( )0, -{a},
since ( )n1 0
a
x , therefore a is a local minimum point.
(a2) If f(n1)(a)0, then from (1) and (3) we have that:
I V x a
f x
f( ) ( )0, ,
therefore a is a local maximum point. Case (b): n+1 is odd:
Here 1
) ( n
a
x has different signs for x < a and for x> a, and from (1) it follows that f(x) f(a) has also different signs in any neighbourhood of the point a, i.e. a is not a local extreme.
x x x
f( )sin ,
! 3 sin
) (
3 x x x x
g ,
x x
x
h( )4 2 32cos2 .
Solution.1. For solving the equation (*): f'(x)0 we have
ZI k k x x
x
f'( )cos 10 k 2 , .
But
"( )sin 0 k
k x
x
f and (3)( )cos 1
k
k x
x
f .
It results that n+1 = 3 is not even; therefore all stationary points for f
are not local extremes. 2. We have:
IR
x x g x x x f x x
x
g ( ) sin ( ) 0,
2 1 cos )
( "
2
' .
It results that g' is a monotone increasing function, therefore a = 0 is a unique stationary point. Moreover
'(0) "(0) (3)(0) (4)(0)0
g g
g
g and (5)(0)10
g .
In conclusion, n+1= 5 is an odd number and g has no local extremes.
3. Since
), 2 ( 4 2 sin 4 8 ) (
'
x f x x
x
h
x=0 is the unique stationary point, '(0) "(0) (3)(0)0
h h
h and (4)(0)320
g ,
so it results that n+1 =4, is even and a = 0 is a local minimum point, and:
hmin=h(0)=3+2=5.
6. ALGEBRAIC APPLICATION
Let Pnbe the linear space of real polynomial functions having
the degree less or equal to nIN, over the field IR, Bc= {1,x,…,xn}
be the canonical basis, and B={1,x-a…(x-a) n}(a 0) another basis
for Pn. If
n
P
f , n
nx
a x
a a x
is a vector written is the basis Bc, since f(n1)(x)0,xIR, then: Using Newton’s binomial formula:
i formula) is:
Exercise 11. Find the inverse of the matrix:
1 0 0 0
3 1 0 0
3 2 1 0
1 1 1 1
T 1 ;
(since: 1 1( 1) 2( 1)2 ...).
x C x
C
xi i i
7. SOLVED PROBLEMS
Exercise 12. Expand the polynomial function:
2 5 5
2 )
(x x x x
f ,
on powers of x1 and x1.
Solution. Since f(6)(x)0 for all xIR we obtain the remainder in Taylor's formula R5(x)=0, for all xIR; therefore:
k k
k
a x k
a f x
T x
f ( )
! ) ( )
( ) (
5
0 ) (
5
where a1, and,in the second case a1. Then: 0
) 1 (
f ; f(1)1122;
x x x x
f
'( )5 4 4 3 4 ,5 ) 1 (
'
f
,f
'(1)5; 412 20
)
( 3 2
''
x x
x
f
,f
''(1)28;f
''(1)12;x x
x
f
'''( )60 2 24 ,84 ) 1 (
'' '
f
;f
'''(1)36; 24120 ) (
) 4 (
x
x
f
,f
(4)(1)144;f
(4)(1)96; !5 ) (
) 5 (
x
f
, hence
1
,1 6 ) 1 ( 14 ) 1 ( 14 ) 1 ( 5 ) 1 ( !
) 1 ( )
(
5 5
0
4 3
2 )
(
x
x x
x x
x k f x
f
k
and
. ) 1 ( ) 1 ( 4
) 1 ( 6 ) 1 ( 6 ) 1 ( 5 2 ) 1 ( !
) 1 ( )
(
5 4
3 2
5
0 ) (
x x
x x
x x
k f x
f k
k k
Exercise 13. Let x
e x x
f( ) 2 . Compute: (a) ( )( )
x f n , for
nIN*. (b) d0f , d0f(x), d f
2
0 , ( )
2 0 f x
d , d03f , 3 ( )
0 f x
d .
(c) d12008f , d12008f(1).
Solution. (a) Using Leibnitz’ formula, for n2, we have that:
2 1
2
2
) ( 2 )
( 2 ) ( 0
) (
) 1 )( 1 ( 2
) 1 ( ) 1 (
) ( ) 1 ( )
( ) ( )
(
n n
n x
n
n k
k x k n k n k
k n x n
k k n n
n n nx x
e
x e C
x e
C x
f
=(1)
2 2 ( 1)
n n nx x
e x
n ,
and we observe that this formula is true for nIN. (b) Generally n n n
a f f a dx
d ( )( ) and n n n
a f x f a x
d ( ) ( ) , for all nIN.
Here:
d0f f'(o)dx0 , d0f(x)0; 2
2 2
0 f f ''(0)dx 2dx
d , d02f(x)2x2;
3 3
3
0 f f' ''(0)dx 6dx
d , 3 3
0 f(x) 6x
d .
(c) 2008 2008 2008
1 f e (1 2 2008 2008 2007)dx
d ,
2008 2008
1 (1) 4026041
e
f
d .
Exercise 14. Using Mac Laurin's formula compute:
x chx
x l
x 2 2
12
lim 2
4
0
)
We observe that n2 verifies this inequality; therefore
Exercise 16. Prove that
n
Solution. The exponential function can be written as a sum of series:
),
whence:
!
and the inequality is proved.
5
0 10
1 ! 1 ! 1
0
k n n
e
n
k
,
and we obtain for 2.71828.
! 1 )
1 ( :
8
8
0
8
k k
T e n
Exercise 17. Write the vector
x xx x x
f( )13 3 2 3IR3
in the base
1, 2,( 2)2,( 2)3
.x x
x B
Solution. Let a2 in Taylor's formula. Then there exists an between 2 and x such that the remainder would be:
0 ) 2 ( ! 4
) ( )
( 4
) 4 (
4 x
f x
R ,
for all xIR; hence:
. ) 2 ( !
) 2 ( )
(
3
0 )
( k
k k
x k f x
f
But:
! 3 '' '
, 6 ) 2 ( '' , 6 6 ) ( ''
; 3 ) 2 ( ' , 3 6 3 ) ( ' ;
3 2
) 2 (
f
f x x
f
f x x x
f f
and
In conclusion: 33( 2)3( 2)2 ( 2)3.
x x
x f
Exercise 18. Let f: I IR→IR, be a function where I is an interval and suppose that there exists ( 1)( )
a
f n IR where aI.
Show that:
), ( )!
1 (
1 )
(
) ( ) (
lim ( 1)
1 f a
m a
x
x T x
f m
m m a
x
n m 1,2,...,
where Tm is Taylor's polynomial of degree m associated to f in the
point a.
). ( )!
1 (
1 )
( 2 ... ) 1 (
) ( )
( lim )
(
) ( ) (
lim ( 1)
) ( )
(
1 f a
m a x m m
a f x f a
x
x T x
f m
m m
a x m m a
x
Exercise 19. Prove the inequality: ,
8 2 1 1
2 x x
x for all x>0.
Solution. Let f :
1,
R, f(x) x1. According to the Taylor's theorem :
0,x so that 2
2 ) ( '' )
0 ( )
(x f x f x
f ,
hence:
1
.8 2 1
1 2
3 2
x x
x
Therefore
, 8 2 1 1
2 x x
x for all x<0.
Exercise 20. Using Taylor's formula prove that:
2 5
cos 4 2 1 lim
2
4 2
0 2
x e
x x chx
x
x .
Solution. Using the well known formulas for elementary functions it follows that there exists c and c’ between 0 and x such
that:
, 5 4
2
! 5 ! 4 ! 2
1 x x shc x
chx
.
Therefore:
)
315 7 15 2 3 ( 315
17 10
) ! 3 ( 5 15
2 3 5 7
7 3
3 4
5 x x x
x x x
x x x
) 10 ) 3 ( 5 ( ! 5
1 ) 9 3 ) 15 2 3 ( 3 ( ! 3
1 3
3 4 5 6
5 3 2 3
x x
x x x
x x
x x
x
=
=
5 3
7
0 15)
2 ! 5 1 ! 3 1 15
2 ! 3
1 ! 5 1 ( ) ! 3
1 3 1 3 1 ! 3 1 ( ) 1 1 [( 1
lim x x x
x
x
) ]
! 4 3
1 6 3
1 15
1 315
17 6 3
2 36
1 15
1 ! 7
1 ! 7 1 315
17
( x7
. 30
1 5 3 2
6 5 3 2
7 5
3 2
1 6
5 3
11 36
3 ! 5 3
2
2 2 2
2 2
2
8.EXERCISES
Exercise 1. Write the following polynomial functions after the powers of x+1 and x+2:
a) ( ) 3 2 1
x x x x f
b) ( ) 4 3 2 5
x x x x f
c) f(x) x5 2x3 3x
Exercise 2. Write the following polynomial functions after the powers of x-1 and y+1:
a) f(x,y) x2 xyy2 x5 b) f(x,y) x3 x2yxyy2 x1 c) f(x,y) x4 y3 xy2 xy
Exercise 3. Using Taylor's formula, show:
Exercise 4. Find the extrema of the following functions: (a) f(x)cosxsinx,x
2 , 2
(b) f(x)sin2xx,
2 , 2
x
(c) f(x)xtanx
(d) f(x)sin3x3sinx
(e) f(x)2xarctanx
(f) f(x)sinxcos2 x
(g)
2 1 , 2 1 ,
2 6 )
( 11 6
x x x x f
(h) ( ) 2 2cos 3
x x
x
f .
Answer. (a) 2
4
max
f
f ;
(b)
6 ,
6 min
max
f f f
f ;
(c) there is neither maximum nor minimum;
(d)
2
min
f
f ,
2 3
max
f
f ; (e) no extrema;
(f)
2
min
f
f ,
3 2 arccos
max f
f , ))
3 2 (arccos(
max f
f ;
(g) fmin f(0)2; (h) fmin f(0)5.
Exercise 5. Find the image of the functions: (a) f :
2,2
IR, ( )3 4 6 2 1x x x
f
(b) f :
1,5
2 3 13 1 ) (
, 3 2
IR f x x x x
(c)
1 1 )
( , 4
, 0 :
x x x f
f IR
(d) f f x xx
, ( ) sin2
2 , 2
(e) 2 4
12 cos 24 ) ( ,
: f x x x x
f IRIR
(f)
3 arccos
) ( , 1
, 1 :
3 x x x x
f
f IR .
Answer. (a)
25,2
; (b)
3 23 , 3 13
; (c)
5 3 ,
1 ; (d)
2 , 2
;
(e)
,24
; (f)
6 5 , 6 7
.
Exercise 6. Prove the inequalities: (a) ,
0,
! 5 ! 3 sin
! 3
5 3 3
x x x x x x
x ;
(b) ,
,
3 2 )
1 ln( 2
3 2 2
o x x x x x x
x ;
(c) 1-x x x x ,xIR !
4 2 1 cos 2
4 2 2
.
Exercise 7. Find the extrema of the following functions:
(a) f f x x x x
3 arctan )
( , :
3
IR
IR ;
(b) , ( ) 2 8
3 1 , 3 1
: 6 3
f x x x
f IR ;
(c) f :
1,1
IR, f(x)x10 2x5 13. Answer: (a),(b),(c) No extrema.Exercise 8. Find the extrema of the following functions:
(a) f :
0,1 IR, f x x x x3 arctan )
(
3
; (b) f :
0,1 IR, ( )2 6 3 8x x x
f ;
(c) f :
o,1 IR, ( ) 10 2 5 13x x x
Answer. (a) fmin f (0)0,
3 2 4 ) 1 (
max
f
f ;
(b)
8 63 4 1
3
min
f
f , fmax f(0)8, fmax f(1)9;
(c) fmax f(0)13, fmin f(1)12.
Exercise 9. Using Taylor's formula prove that: (a) lim 34
2 cos( 2 )
10
x e x
x
x ;
(b)
12ln 3 16 36 48 25
1) 1 ( 12
5
lim 5 4 3 2
1
x x x x x x
x ;
(c)
( 6 ) 6 ( 5 ) 15 ( 4 ) 20 ( 3 )
1 lim 6
0h f x h f x h f x h f x h
h
( ), )( ) ( 6 ) 2 (
15 (6)
x f x f h x f h x
f
where f :IRIR and f (6) on IR.
(d) lim 23
tan sin
10
x x x
x ;
(e) lim2 2ln 1 1
x
x x x
x ;
(f)
ln(1 2 ) 2 sin2
1 41
lim 3 2
0
x x x x
x ;
(g) 1
) 2 2
( 6
cos )
sin( lim
2 2
0
xshx chx
x x x
x ;
(h) sin 1
3 2
! 7 lim
3 7
0
x
x shx x
x .
Exercise 10*. Let f :IRIR having the second derivative on IR, such that f(x)13, f'(x) 13 for all xIR. Prove that
26 ) ( ' x