Vous êtes sur la page 1sur 21

Index FAQ

1. POWER SERIES
2. TAYLOR POLYNOMIALS AT X = 0
3. BASIC TAYLOR POLYNOMIALS
4. TAYLOR SERIES AND MACLAURIN SERIES
5. TAYLOR POLYNOMIALS AT X = A
6. FINDING TAYLOR SERIES BY SUBSTITUTIONS,
DIFFERENTIATION AND INTEGRATION
7. APPROXIMATION OF FUNCTIONS

Taylor and Maclaurin Series
An important result from calculus used in
Numerical Analysis
Dr T Murugesan 1
Index FAQ
Dr T Murugesan
Functions Represented by Power Series
( ) ( )
( ) ( ) ( )
( )
( )

=
' '' '''
= + + + + =

2 3
0
f 0 f 0 f 0 f 0
f f 0 .
1! 2! 3! !
k
k
k
x x x x x
k
( )
( )
= =
f 0
Insert 0 to get .
!
k
k
x a
k
( )
( )
( ) ( )
+ +
= + + +
th
2
The general derivative of f is
1 ! 2 !
f ! .
1! 2!
k
k
k
k k
x k a x x
Consider a function f represented by a converging power series:
f(x) = a
0
+ a
1
x + a
2
x
2
+ a
3
x
3
+ .
Conclude
Clearly f(0) = a
0
.
To compute the other coefficients a
k
, differentiate term by term to get
f(x) = a
1
+ 2a
2
x + . Insert x = 0 to the above to get f(0) = a
1
.
A converging power series representing a function f is necessarily of
the above form. This power series is called Maclaurin Series, a
special form a more general Taylor Series.
By inserting x=0.
2
Index FAQ
Dr T Murugesan
Approximating Functions
Another way to get to Taylor Series is to consider
approximations of functions by polynomials.
Definition
The Taylor polynomial of degree n for given
function f(x) at a point a is a polynomial P(x) of
degree n such that
P
(k)
(a)=f
(k)
(a) for k=0,1,,n.

This means that the value of the polynomial P(x) and all of its
derivatives up to the order n agree with those of the function f (x)
at the point x=a.
Observe that the defining conditions for the Taylor polynomial have
to do with the behavior of the polynomial at one point only.
We assume that the continuous function f has derivatives of all
orders everywhere in its domain of definition.
3
Index FAQ
Dr T Murugesan
Taylor Polynomials at x = 0
0 1
Let P be a Taylor polynomial for a given
function f at 0. Assume that f has derivatives of arbitrary high
order at 0.
n
n
a a x a x
x
x
= + + +
=
=
0
P(0) a =
2
P''(0) 2a =
1
P'(0) a =
3
P'''(0) 3 2a =
( )
P (0) ( 1) 3 2 !
k
k k
k k a k a = =
Straightforward differentiation yields.
The general formula is
( )
( ) ( )
f (0)
The defining condition P (0) f (0) yields: .
!
k
k k
k
a
k
= =
We conclude that the Taylor Polynomial for an infinitely
differentiable function f (x) at x=0 is uniquely defined, and
that the coefficients a
k
are given by the above formula.
4
Index FAQ
Dr T Murugesan
Explicit Taylor Polynomials
( ) ( )
( ) ( ) ( )
( )
( )
' '' '''
= + + + + +
2 3
f,
f 0 f 0 f 0 f 0
T f 0
1! 2! 3! !
n
n
n
x x x x x
n
Formula
Using this formula Taylor polynomials of functions can often be
rather easily computed. Strategy is the following:
1. Compute several derivatives of the given function.
2. Evaluate these derivatives at x = 0.
3. Detect a pattern to find a general formula for f
(n)
(0).
By the preceding considerations, the Taylor
Polynomial of degree n of a function f is the
polynomial T
fn
(x) or P
n
(x)
5
Index FAQ
Dr T Murugesan
Taylor Polynomials for the Sine Function
( )
( )
+
+
+
=

= + + =
+ +

2 1
3 5 2 1
sin,2 1
0
1
1 1 ( 1)
T .
3! 5! (2 1)! 2 1 !
k
n n
n
k
n
k
x
x x x x
n k
Formula
( )
( )
( )
( ) ( )
+
= = =
2 2 1
f 0 0 and f 0 1 for all 0,1,2, .
k
k k
k
( )
( ) ( )
( )
( )
= =
4 4
f sin , f 0 0. x x
( )
( )
( ) ( ) ( )
= = =
0
f f sin , f 0 0. x x x
To find the Taylor polynomials of the function f(x) = sin(x) compute
derivatives and evaluate them at x = 0. One gets:
( )
( )
( ) ( ) ( )
' ' = = =
1
f f cos , f 0 1. x x x
( )
( )
( ) ( ) ( )
''' ''' = = =
3
f f cos , f 0 1. x x x
( )
( )
( ) ( ) ( )
'' '' = = =
2
f f sin , f 0 0. x x x
Note:
All even order derivatives
of sin take the value 0 at x
= 0. Odd order
derivatives take the values
1 and -1.
Conclusion as a Formula
6
Index FAQ
Dr T Murugesan
Taylor Polynomials for Cosine
( )
( )
+
=

= + + =

2
2 4 2
cos,2 1
0
1
1 1 ( 1)
T 1 .
2! 4! (2 )! 2 !
k
n n
n
k
n
k
x
x x x
n k
Formula
( )
( ) ( )
( )
( )
+
= = =
2 2 1
g 0 1 and g 0 0 for all 0,1,2, .
k
k k
k
( )
( ) ( )
( )
( )
= =
4 4
g cos , g 0 1. x x
( )
( )
( ) ( ) ( )
= = =
0
g g cos , g 0 1. x x x
To find the Taylor polynomials of the function g(x) = cos(x)
compute derivatives and evaluate them at x = 0. One gets:
( )
( )
( ) ( ) ( )
' ' = = =
1
g g sin , g 0 0. x x x
( )
( )
( ) ( ) ( )
''' ''' = = =
3
g g sin , g 0 0. x x x
( )
( )
( ) ( ) ( )
'' '' = = =
2
g g cos , g 0 1. x x x
Conclude
All odd order derivatives
of cos(x) take the value 0
at x = 0.
Even order derivatives
take the values 1 and -1.
Conclusion as a Formula
7
Index FAQ
Dr T Murugesan
Taylor Polynomials for the Exponential
Function
+
=
= + + + + =

2
exp,2 1
0
T 1 .
1! 2! ! !
n k
n
n
k
x x x x
n k
Formula
( )
( )
( ) ( )
= = =
0
h h e , h 0 1.
x
x x
To find the Taylor polynomials of the exponential function
h(x) = e
x
compute derivatives and evaluate them at x = 0.
One gets:
( )
( )
( ) ( )
' ' = = =
1
h h e , h 0 1.
x
x x
( )
( )
( )
( )
= = h e , h 0 1.
n n
x
x
Note:
The value of the exponential
function and that of all of its
derivatives at x = 0 is 1.
8
Index FAQ
Dr T Murugesan
Taylor Polynomials for the Sine Function
+
+

= + +
+
2 1
3 5
sin,2 1
1 1 ( 1)
T .
3! 5! (2 1)!
n n
n
x
x x x
n
Formula
The following figure illustrates Taylor polynomials of degrees 5
(blue), 9 (red) and 15 (green) for the sine function.
One concludes from the picture that all of the above Taylor
approximations for the sine function appear to approximate the
function well near the origin (center of the above picture). Higher
order Taylor polynomials approximate better away from the
origin.
9
Index FAQ
Dr T Murugesan
Basic Taylor Polynomials
( )
=

= + + =

2 4 2 2
cos,2
0
1 1 ( 1) ( 1)
T 1
2! 4! (2 )! (2 )!
n k
n
n k
k
k
x x x x x
n k
We will later see that the above polynomials can be used to
approximate the values of the respective functions for all x.
1
( )
+ +
+
=

= + + =
+ +

2 1 2 1
3 5
sin,2 1
0
1 1 ( 1) ( 1)
T .
3! 5! (2 1)! (2 1)!
n n k k n
k
k
x x
x x x x
n k
2
( )
=
= + + + + + =

2 3
exp,
0
1 1 1 1
T 1
2! 3! ! !
n
n k
n
k
x x x x x x
n k
3
( )

= + + + + +
2 3
f,
( 1) ( 1)( 2) ( 1) ( ( 1))
T 1
2! 3! !
n
n
p p p p p p p p n
x px x x x
n
4
The Taylor Polynomials of the function f(x) = (1+x)
p
are given
below. They can be used to approximate the values of the function
only for -1<x<1.
10
Index FAQ
Dr T Murugesan
Taylor Polynomials at x = a
The conditions used to define a Taylor polynomial P of a given
function f require that the polynomial P and all of its non-zero
derivatives at a point x=a agree with those of the function f. Clearly
the definition implies that the polynomial P approximates the
function f best near the point x=a.
Formula for Taylor Polynomials at
x=a
Assume the function f has
all derivatives at the point
x=a.
( )
2
f( ) f( ) f ( )
P( ) f( ) ( ) ( ) ( )
1! 2! !
n
n
a a a
x a x a x a x a
n
= + + + +
Taylor polynomial of degree n at x=a
is
The above formula follows by directly computing the values of the
derivatives of the function f and those of the polynomial P at
x=a.
11
Index FAQ
Dr T Murugesan
Goodness of Approximations
The following figure shows the graph of the sine functions and those
of its Taylor polynomials of degree 5 at the points x=-t, x=0, x=t,
x=2t, x=3t.
The Taylor polynomial of degree 5 at the point x=-t approximates
the sine function so well near the point x=-t that its graphs is
completely covered by the black graph of the sine function near that
point. As x< -3t/2 or x> -t/2 the approximation fails to follow the
graph of the sine function. These portions of the graph of the Taylor
polynomial of degree 5 at x=-t are shown as the left most red
graphs above and under the x -axis.
12
Index FAQ
Dr T Murugesan
Taylor Series and Maclaurin series
2 3
( 1) ( 1)( 2)
(1 ) 1
2! 3!
p
p p p p p
x px x x

+ = + + + +
Letting n grow the Taylor polynomials at x = a define Taylor series at x
= a for the respective functions. Basic Taylor Polynomials yield the
following Basic Taylor Series at x = 0. The Taylor Series at x = 0 are
also called Maclaurin series.
These series
converge
and
represent
the given
function for
all x. This
will be
shown later.
This is
valid for
-1<x<1.
The Binomial Series
2 4 2
0
1 1 ( 1)
cos( ) 1
2! 4! (2 )!
k
k
k
x x x x
k

= + =

1
2 1
3 5
0
1 1 ( 1)
sin( ) .
3! 5! (2 1)!
k k
k
x
x x x x
k
+

= + =
+

=
= + + + + =

2 3
0
1 1 1
e 1
2! 3! !
x k
k
x x x x
k
3
13
Index FAQ
Dr T Murugesan
Taylor Polynomial Approximations
( ) | |
~ +
3 5
Estimate the error of the approximation
1 1
sin in the interval -1,1 .
3! 5!
x x x x
Problem :1
( ) ( )
= + = +
7 9
3 5
1 1
E sin
3! 5! 7! 9!
x x
x x x x x
Solution
( )
s s = ~ s s
7
1 1
E 0.000198 for 1 1.
7! 7! 5040
x
x x
The error E(x) is given by F(x) P(x) , hence
The Taylor series for sin(x) is an alternating series.
The error done when approximating sin(x) by a Taylor
polynomial of degree 5 P
5
(x)is bounded by the absolute value
of the first term left out. Hence
14
Index FAQ
Dr T Murugesan
Binomial Series and Geometric Series
2 3
( 1) ( 1)( 2)
(1 ) 1
2! 3!
p
p p p p p
x px x x

+ = + + + +
The Binomial Series
( )( ) ( )( )( )

= + + + = + +
+
2 3 2 3
1 2 1 2 3
1
1 1 .
1 2! 3!
x x x x x x
x
Insert p = -1 to the above to get
Conclude The binomial series for p = -1, is the geometric series
with the first term 1 and with q = x as the ratio of two
subsequent terms.
15
Index FAQ
Dr T Murugesan
Binomial Series Example

+ = + + + +
2 3
( 1) ( 1)( 2)
(1 ) 1
2! 3!
p
p p p p p
z pz z z
Use the Binomial Series
( )
( )( )
( )( )
( )
( )( )( )
( )

= + + + +

2 3
2
2 3 2 3 4
1
1 2
2! 3!
1
x x x
x
Insert p = -2 to the above to get
Problem: 2 Find Taylor series for the function f(x) = 1/(1 - x)
2.
( )
( )

=
= + + + + = +


3
2
0
1
1 2 3 4 1
1
k
k
x x x k x
x
Alternative derivation
for this series
expansion follows later.
These terms cancel.
16
Index FAQ
Dr T Murugesan
Finding Taylor Series
One can find Taylor series for complicated functions by
1. Substitutions
2. Integrating a known series term by term
3. Differentiating a known series term by term
4. Any combination of the above tricks
One usually starts with one of the basic Taylor series and
manipulates that to get the desired Taylor series. The above tricks
are legal provided that the series in question converge and
represent the functions in question. This depends on the function
for which Taylor series representation needs to be derived. Many
of the basic Taylor series converge everywhere.
17
Index FAQ
Dr T Murugesan
Finding Taylor Series by Substitution

=
= + + + + =

2 3
0
Start with the basic Taylor series
1 1
e 1 .
2! 3! !
n
z
n
z
z z z
n
2
2
2 4 6
0
One gets
1 1 ( 1)
e 1
2! 3! !
n n
x
n
x
x x x
n

= + + =

( )
2
Find Taylor series for f e .
x
x

=
Problem: 3
2
Substitute - . z x =
Solution
Suggest some more uses for exponential function
18
Index FAQ
Dr T Murugesan
Finding Taylor Series by Integration
( ) ( )
2
0
1
Hence f 1 1 .
1
n
n
n
x x x x
x

=
' = = + =
+

( )
( )
1
0 0
1
1
1 .
1 1
n
n
n n
n n
dx x dx x C
x n

+
= =

| |
= = +
|
+ +
\ .

} }
( )
( )
1 2 3
0
1
1 1
This implies ln 1 .
1 2 3
n
n
n
x x C C x x x
n

+
=

+ = + = + +
+

Evaluating the above for 0 one gets 0. x C = =


Problem: 4
( ) ( )
Find Taylor series for f ln 1 . x x = +
Solution
( )
1
Observe that for this function f .
1
x
x
'
=
+
Can you suggest one more example?
19
Index FAQ
Dr T Murugesan
Taylor Series for ln(1 + x)
( )
( )
1
2 3 4
1
1
1 1 1
ln 1 .
2 3 4
n
n
n
x x x x x x
n
+

+ = + + =

Formula
This figure shows the graph of the
function ln(1 + x) and those of its
Taylor polynomials of order 4 (blue),
9 (red), 14 (green) and 19 (yellow).
1
1
4
4
1
9
9
One observes that up to x = 1
higher order Taylor polynomials
give better approximations than
lower order Taylor polynomials. For
x > 1, the situation is reversed: the
higher the order, the worse the
approximation.
This reflects the fact that the
Taylor series for ln(1 + x)
does not converge for x > 1.
20
Index FAQ
Dr T Murugesan
Finding Taylor Series by Differentiation
2
0
1
and recall that 1 .
1
n
n
x x x
x

=
= + + + =


2
1 1
Observe that ,
(1 ) 1
d
x dx x
| |
=
|

\ .
The above formula is a special case of the binomial series
and it converges for |x| < 1.
Problem: 5
( )
2
1
Find Taylor series for f .
(1 )
x
x
=

Solution
1 2
2
0 0
1 1
Hence 1 2 3
(1 ) 1
n n
n n
d d
x nx x x
x dx x dx

= =
| | | |
= = = = + + +
| |

\ . \ .

21

Vous aimerez peut-être aussi