Anda di halaman 1dari 9

Taylor series and analytic functions

The infinite, the series (Taylor) and the analytic functions .



George Mpante mathematics teacher
mpantes_2@yahoo.gr
introduction
numerical series
The sum and the limit
The history of series
Series of functions
Juethadera,
Taylor series, analytic functions
Some results
omment
!ntroduction .
Infinity is omnipresent in mathematics agenda, hidden or apparent, in all eras. We
know how much attention and perspicacity Greeks handled the concept of irrational number,
associated with the idea of infinity. But even in the simplest numbers, the natural, the infinity is
behind their unlimited succession , in rational numbers behind their decimal form.
The first contact of mathematical practice with infinity was in the infinite sums of
numbers - now called series- which occurred through direct uestions about the continuity of
concepts, such as space, or time , i.e if a distance can be divided in two, four, eight .....ad
infinitum parts, where it is easy to put aside reality and renegotiate the uestion in terms of
abstract ob!ects which are infinitely divisible" the numbers #the thin"a#le ghost #ehind
phenomena, $ristotle) .
%umerical series .
1
George Mpantes mathematics teacher
$uppose we take the algebraic sum of one million finite numbers. The sum is clearly
finite . But if we assume that we consider an infinite succession of numbers, we have an
infinite series . %ow we can not speak of the sum of the terms since infinity is never
e&hausted. We can not speakof the sum of an infinite series , since infinity is never a&hausted
. But we may consider magnitude which plays the same role with respect to an infinite series
as the sum does to a finite succession of terms. This magnitude is the one towards which the
sum of the first n terms of the infinite series tends when n is made to increase indefinitely, It
is a non-algebraic concept, called the limit of the infinite series.
&'ample ( .
We shall find the limit of the series
' () * + () * , () * )- (,,,,,,,,,,,
. is the sum of the first n terms
. / )*0 () * + () * , ( ... )*0
n
#)1 which will calculate in algebra.
. / )-)*0
n
22.#01
3ssuming now that n increases infinitely, the fraction tends to 4ero and have the limit of .,
the limit $ / ), the series converges, the limit number is finite.
The sum and the limit .
Important differences distinguish the sum of a finite number of numbers from the
limit of an infinite number of numbers. 5et us mention a few.
We 6now that the sum of a billion numbers, some of whish are positive and others
negative has e&actly the same value whether we effect the summation
in the order given or in any other order. But as was proved by 7irichlet
in the nineteenth century , this conclusion is not necessary correct
when we are dealing with infinite series" the limit may have one value
or another according to the way in which the terms are ordered.
But there are and other distinctions between the sum and the
limit. Thus the sum of a finite number of numbers is always finite and ,
as !ust e&plained, is perfectly well determined. 8n the other hand, the
limit of an infinite series may be finite or infinite, , and sometimes the
series has no limit. In the two latter cases the series is said to di)erge .
9or e&ample, if we have the unending series ) () () () ( ........ the limit is infinite ,
the series diverges .
$imilarly , if we have a finite seuence of terms such as
).) ()-) () or )-)()-)()-)
the sum will have the value : or ) according to whether the number of terms is odd or even.
But if we are dealing with an infinite series of this type , there is no definite limit , for we have
a kind of oscillation between : and ) ; the series diverges.
2
Taylor series and analytic functions
3s an illustration of a series the limit of which is well defined and finite, we may
mention the series
)()*0()*+()*,(2.. its limit is 0. The series
)-)*0()*<-)*+()*=-2.. also converges, but here we must be careful not to change the
order in which the terms are written. If we change the order of the terms ,we may find that the
series diverges.
Today with the notation
we mean the series
and with the notation
we mean the limit of series.
The *istory of series .
The first series appeared is that of +eno #e&ample )1
' () * + () * , () * )- (22
which caused the historical parado& of dichotomy.
What was the source of the parado&>
?eno saw the limit as a sum within the @ythagorean culture of distinctness, he did not
conceive the meaning of limit, #here it was ), but for ?eno it was infinite as we donAt stop
adding terms1 he did not understand infinity through
mathematics. 3 cavalier treatment of infinite can lead to
absurd resultsB
The first who historically calculated a limit of a series
was $rchimedes. Ce proved that the area of a parabolic
segment 3BD7E is +*< the triangle 3DE having the same
base and verte&. 9or this purpose, he constructed
continuously an unlimited seuence of inscribed triangles,
starting from the initial 3DE of area &, adding triangles
between the e&isting. Ce conceived that the area of the
parabolic segment will be covered after infinite such triangles,
and by geometrical tricks, #FGH IJK LMJNMIOPQNMKJK1, he did not the concept of limit, he
found that
&#)( )*+( )*)-( )*-+(222.1/+*<& or
)( )*+( )*)-( )*-+(222.1/+*<
)

1
. These euations are no less certain than the others2albeit we Rortals whose
reasoning powers are confined within narrow limits , can neither e&press nor so conceive all
the terms of these euations , as to 6now e&actly from these the uantities we
3
George Mpantes mathematics teacher
This process was a conceivable experiment in the field of geometric intuition. The
experiment does not end but we watch its course as we want, even we imagine this. Hence
we perceive the result of the experimen.. 5ater, this geometrical process was replaced by
numbers, and perfected when we completed our knowledge about the real numbers
Sntil the ),th century the procedures of infinity were treated without rigor and
precision , perhaps as reported by $truik, because of uncontrolled enthusiasm # history of
mathematics 7irk $truik 1 .
6nown limits of series in the period before Dauchy were
#5eibni41 and
#Euler1
$o maddening infinity initially disappears thanks to our own
ingenuity , and the results were strange . 23s %ewton , 5eibni4, the several Bernoullis,
Euler , dA 3lembert, 5agrange, and other ),
th
-century men struggled with the strange problem
of infinite series and employed them in analysis, they perpetrated all sorts of blunders , made
false proofs, and drew incorrect conclusionsT they even gave arguments that now with
hindsight we are obliged to call ludicrous2.. #Rorris 6lein 1.
U ..... But questions linger centuries answered , preparing us for a long and
deep mystery .U
auchy, # ultimately in )Vth century 1 observed that between the infinite and the
infinitesimal there is an interdependence . If we look at the seuence ),) * 0, )*<, W, )*=, .......
will observe that with increasing the order of terms, they are closer to 4ero . 3pproaching
4ero does not end # infinitesimal 1 as does not end the succession , #infinite 1 U ... when the
successive values given in a variable, approach indefinitely a fixed value differing from it as
one wishes, the latter is called limit of all others .... U noted Dauchy and translated this
definition in a rigid formulation "
,efinition (. say that the number X is the limit of the seuence &n if for every e Y :
there is n: Z% such as for n Y n: we have [ &n - X [ \ e. The se-uence con)erges to . and
we now have a strict convergence criterion for all seuences.
Based on this definition, a calculus of limits developed for series, #root test of
Dauchy, ratio test of 7A3lembert, comparison test ..1. These tests stopped the tricks for each
series separately, we can manipulate satisfactorily many problems in the convergence of
series, that up Euler was the weak point of mathematics.
Series of functions .
$eries of functions differ from numerical series in that the successive terms are
functions of a magnitude & whose precise numerical value is not specified. Thus the infinite
series
want2..%e/ton
4
Taylor series and analytic functions
f)#&1(f0#&1(f<#&1 (222..
where f)#&1, f0#&1, f<#&1,2..are known functions of &, represents a series of functions.
The functions now define numbers, and our series of functions becomes a numerical series.
By changing the value ascribed to & , we define one numerical series after another. 9or
e&ample a series of functions is
&-&
<
*<(&
=
*=-&
]
*](22..
for &/) we have the numerical series
)-)*<()*=-)*](22.
9or &/0 we have
0-,*<(<0*=-)0,*](22..
In most cases the numerical series converges if certain values of & are substituted,
whereas it diverges #i.e. has no limit or becomes infinite1 when other values of & are taken.
$uppose for argumentAs sake that the series converges for all values of & of the interval
#a,b1, and diverges for all other value of &. The limit of the series then has a welll-determined
value only when & is situated between a and b, and usually this value will depend on the
value assigned to & in the interval. The limit of the series thus defines a function of &. But note
that it is only when x is situated between a and b (where the series converges) that the series
defines a function; elsewhere the series is meaningless and defines nothing.

Jyesthadera .
3 first occurrence of the phenomenon of series of functions comes from India, one
hundred years before the results of the Europeans. The Indian Jyesthadera #)=:: 37 1
e&pressed the number ^ as a numerical series. @reviously demonstrated algebraically the
e&pansion"

The trick is algebraic and piecemeal, we know nothing deeper. Ce replaces the parenthesis
2
In ),0, 3bel wrote" divergent series are the invention of the devil and is unorthodo& to rely
on them to prove anything.
5
George Mpantes mathematics teacher
from the identity"

_yesthadeva thereby turned a simple fraction into an infinite series.
But he did not suspect that the effects of formula 0 apply for certain values of &. If we set & / )
in both sides we have that
' / )-) ()-) ()-) ........ which is not true as stated .
9or & / -) have the result )*: / `
and & / )*0 we have
0*< / )-)*0 ()*+-) * , () * )- ....... which may be true or not. I.e. 9or different values of & the
result of _yesthaderan leads to Utruth or mystery or nonsense.U #7avid !er"ins1
Today we know that the formula # 0 1 is valid for -) \ & \), he could not know this, but
in his proof, the geometric entanglement of the formula in a figure, ensured incidentally
that ^ is in the above range. $o the formula was correct for &/^ although _yesthadera did
not note it.
With formula #0 1 and geometry _yesthadera proved the surprising result in the
numerical series "
#<1
a2.#ho would suspect this lin" between the constant $ and a series involving the
odd numbers>..... %avid !er"insb.
Taylor series , analytical functions .
This process of _yesthaderan was the task of life of Brook Taylor # www.mpantes.gr,
&' ()*+(&,-./)'0 *10 /234,*1/101, who calculated the necessary and sufficient condition that
a function be developed in power series, ie a polynomial with infinite terms #long polynomial1,
that is a technical and difficult sub!ect , the known as Taylor series.
6
Taylor series and analytic functions
If f has infinitely many derivatives at the point where &/a, we conclude with Taylor that
the power series generated by f at &/a is
) 4 .......( .......... ) (
!
) (
) (
0
) (
k
k
k
a x
k
a f
x f =

=
But the second half of the transformation was to study the convergence of the
resulting series . This was ignored by Taylor.
The complete study of power series was by the founders of the comple& analysis
#Dauchy, ciemann, 3bel, Weierstrass1 a century after their construction.
The series #+1always converges for the value &/a since it reduces to the first term.
But the power series may also converge for some other value of &. Rathematical analysis
shows that in this case the series will necessary converge for a continuous range of
values of & and that the point &/a occupy the centre of the range. Within this range of
convergence, the power series defines a function of & which can be shown to be a continuous
function, all of whose derivatives are also continuous. 3 power series of this last kind #which
converge in an interval1 constitutes a Taylor series so that a continuous function defined by
a power series is what we have called an analytic function of '.
&'ample0 the function y/sin& 22#=1
$ubstituting the formula #+1 for the coefficients in Taylor e&pansion we have
) 6 .....( ..........
! 7 ! 5 ! 3
7 5 3
+ + =
x x x
x y
det we canAt euate the right sides of #=1 and #-1 before we study the convergence of
the series. 9or some & could the two members give different resultsB Today we conclude
#ratio test1 that the Taylor series in #-1 converges for all & Zc, that is the domain of the
function. %ow the two e&pressions #=1 and #-1 are euivalent and the function y/sin& is
called the analytic function. Taylor considered it for granted for all the functions. But all
functions known to the mathematicians of the ),
th
century happened to be analytic. Thanks to
this coincidence, #analytical 1 function became a mathematical tool capable highlights the
corollaries of mechanical motions for point masses and rigid bodies and solve the differential
euations describing the laws of these phenomena . $o the successful resolution of many
problems in physics and celestial mechanics, which would not happen without the Taylor
series, due to the fact that the functions involved in these areas were analytic . The c o r r e c
t r e s u l t s w e r e i n c i d e n t a l because the convergence conditions were applied to
the problem, unnoticed by the scholars .
It is really a fact that the progress of physical theories is dependent of the mathematical
development of the day. But while the rigor is essential for mathematics is often preferable
not to reuire in early stages. 3n early insistence on rigor could strangle discovery. Cistory
7
George Mpantes mathematics teacher
shows that the details can wait ... .. Uwhat serves our admiration for the work of the builder if
we can not appreciate the design of the architect> 1oincare e
Some results
The analytical function described relations #there are many such in the world of
engineering1 which are not directly e&pressed with elementary e&pressions but with
successive appro&imations, with both practical results #however accurate 1 getting some
terms of the series, also theoretical results by resorting to limit. In this sense the work of
Brook Taylor is one of the most important works of Rathematics.
The e&tension of the capabilities of the function after itAs development in Taylor
series #analytical function1 is obvious.
There was no way to e&press the initial of y/ sin&*& or
2
x
e y

=
There were computational problems for e&ample the value of cos)
:
3s cos&/)-&
0
*0(&
+
*+-&
-
*-(2. and for &/^*),: from the two first terms we have
cos)
:
/)-:,:::)=/:,VVV,=.
We can find appro&imately the solution of the euation cos&-0&
0
with the previous
series or the indefinite integrals of sin&*& , or e
-&0
, as and differentiation of functions, because
technically it is easier to work with the simpler terms of the series.
But the more substantial contribution of analytic functions was to solve differential
euations with great practical and theoretical importance. The euations of 3iry, 5egendre,
Cermitte, Bessel, are euations that their solutions can not be written in terms of known
functions such as polynomials, e&ponential or trigonometric, but only be e&pressed in a
power series. 9or e&ample, the differential euation of the pendulum could not be resolved,
the physical description were at an impasse .
omment
The graph of an analytic curve called analytical curve. This has an almost
metaphysical status. The mathematical analysis shows that if we know e&actly a part of it, no
matter how small, we can design the whole, we know the whole routeB If two analytical curves
coincide at a small fraction of them arbitrarily then necessarily coincide in their entire length.
Cere we can diagnose the doctrine of causality in physical systems. If we know the initial
conditions we can describe their evolution. That is described here is the uncanny ability of
mathematics to describe nature, analytical curves involved in the doctrine of causality of
natural systems.
George Mpantes
mpantes_2@yahoo.gr ///. mpantes.gr . .
8
Taylor series and analytic functions
Sources .
fG NMIHNPOghiMGj Ikj ilKmOIkikj nGhOLPj o^HKIpj #www .mpantes .gr1
Rodern mathematical analysis Rurrey @rotter #3ddison-Wesley publishing Dompany1
oHqkNHIGrm s tlrMuPl #fvws1 #H^x IH rHXQIMOH yGyXuH iIP tQrMGP1
3 short account of the history of Rathematics" W.W. couse Ball
Dalculus and its origins , 7avid @erkins #Rathematical association of 3merica1
nMKGrm oHqkNHIGrm #zxNPG {, s1 |. {KHiIHiGmFkj, }Mi*Kurk
}MJOuH oGLHFGrhK .lKHOI~iMJK |. fGrPKPNuFkj, }Mi*Kurk
.zfv{ {|zv| o{}o{z| vof.ov||, .z. s )V--
The Taylor $eries, by @. 7ienes, 7over @ublications
Infinite seuences and series by, 6onrad 6nopp, 7over @ublications
3 short account of the history of Rathematics by, W.W.couse Ball , 7over @ublications
Rathematics and the physical world, by, Rorris 6line, 7over @ublications
Rathematics the loss of certainty by Rorris 6line, 7over
.lKP^IGr~ iIPOuH IJK oHqkNHIGrhK 7irk $truik wHuFHXPj, HHOx^PlXPj
The rise of the new physics , by 3.7A 3bro , 7over
9

Anda mungkin juga menyukai