You are on page 1of 2

Natl. Acad. Sci. Lett.

(May–June 2017) 40(3):215–216


DOI 10.1007/s40009-016-0527-0

SHORT COMMUNICATION

Relation Between Fourier and Taylor Series


A. Guha1,2

Received: 17 October 2015 / Revised: 7 July 2016 / Accepted: 5 December 2016 / Published online: 19 January 2017
Ó The Author(s) 2017. This article is published with open access at Springerlink.com

Abstract Infinite series can converge in various ways to The foundation of Fourier series1 is based on the principle
give the resultant function. Particularly, here, we consider of orthogonality of functions. Now, the coefficients for the
the Fourier series and compare it with its Taylor equivalent Fourier Series can easily be derived by using the orthog-
both of which are convergent infinite series in their own onality condition. If we try to apply a similar analogy to
rights. However, these are valid under separate limiting Taylor Series coefficients, then the function which obtains
conditions. We consider what happens if we try to derive all the Taylor coefficients turns out to be not quite a
one series from the other or see if such a derivation is function, but a distribution called the Dirac delta function.
possible at all and its implications. An expansion of the This will explicitly be used later to derive one from the
Dirichlet kernel, while using a form of the Dirac delta other later.
function has been shown to yield the Taylor series in its The Fourier series can be succinctly represented in the
form. However, it introduced certain restrictions on both form of the Dirichlet kernel as:
the local and global nature of such a function. Z
1 p
f ðxÞ ¼ lim f ðtÞRrn¼r einðxtÞ dt ð1Þ
r!1 2p p
Keywords Fourier series  Taylor series  Dirichlet kernel  Z
Dirac delta function  Heaviside step function 1 p sinð2r þ 1Þ xt2
¼ lim f ðtÞ ð2Þ
r!1 2p p 2 sin xt
2

Using a certain representation of the Dirac delta function


that suits us:
Z
1 n ixt
dn ðxÞ ¼ e dt
2p n
ð3Þ
sin nx
¼
px
Therefore, as a slight modification, the Dirac delta can be
represented (in the limit of large n) as:
1 sinðn þ 1=2Þx
dn ðxÞ ¼ ð4Þ
2p sin 1=2x
Introducing a change of variables with s ¼ x  t, we solve
& A. Guha
ag922@cam.ac.uk the above integral by repeatedly using by parts. Performing
the first 2 iterations gives us:
1
Department of Physics, St. Stephen’s College, University of
Delhi, New Delhi, India
2
Institute of Astronomy, University of Cambridge, Cambridge,
1
UK If the function satisfies the Dirichlet conditions.

123
216 A. Guha

 Z px
1 sinðr þ 1=2Þs Thus going on expanding the last term by parts over and
f ðxÞ ¼ lim f ðx þ sÞ ds over, we get:
r!1 2p px sin s=2
Z px Z   ð5Þ 00
0 sinðr þ 1=2Þs 0 f ðx þ sÞðp  xÞ2
 f ðx þ sÞ ds ds f ðxÞ ¼ f ðx þ sÞ  f ðx þ sÞðp  xÞ þ
px sin s=2 2!
000 0000

The integration of the Dirac delta function gives the f ðx þ sÞðp  xÞ3 f ðx þ sÞðp  xÞ4
 þ  
Heaviside step function where it is defined as:2 3! 4!
Z x ð9Þ
hðxÞ  dðsÞds ð6Þ
1 Transforming:
A notation is thus introduced here so that we can suffi- s :¼ ðp  xÞ ¼ h
ciently represent the successive integrals of the Heaviside
We thus obtain:
function as we expand similarly the other terms, continu- 0 00
ally invoking by parts. h0 ðxÞ is the Heaviside function hf ðx  hÞ h2 f ðx  hÞ ð10Þ
f ðxÞ ¼ f ðx  hÞ þ þ þ 
while h1 ðxÞ, h2 ðxÞ; . . .; hn ðxÞ represent the first, second, . . . 1! 2!
and nth integrals of the Heaviside function respectively. Specifically:
Because of the nature of the Heaviside function, we note 0 00
immediately that the function with successive integrations hf ðpÞ h2 f ðpÞ
f ðp þ hÞ ¼ f ðpÞ þ þ þ  ð11Þ
performed will only resemble the function in the interval 1! 2!
p to p. Thus after a few iterations, we obtain: This shows us that the value of the function and its
Z px
0 derivatives at p is extremely important. This can then be
f ðxÞ ¼ f ðx þ sÞ  f ðx þ sÞ h0 ðsÞds further generalised to a Fourier series with any suitably
px
Z px well-defined period. Thus the Taylor series is shown to be
00
þ f ðx þ sÞ h1 ðsÞds ð7Þ exactly derivable from its Fourier series. However there are
Zpx
px a few important caveats which are listed below.
000
 f ðx þ sÞ h2 ðsÞds þ    It must be noted that the above formulation of equality
px
holds only for the specific transformation relation. Gener-
Now, the integrations of these class of h functions can be alizing, this shows that the global properties of Fourier
performed trivially to obtain as follows: series can be matched by the local properties of expanding
a function using a certain expansion coefficient. This above
h1 ðxÞ ¼ xh0 ðxÞ
formulation could be replaced for a general Fourier series
x2 with the period replacing the limits p to p.
h2 ðxÞ ¼ h0 ðxÞ ð8Þ
2 Thus, a neat relation has been shown to exist between a
... function’s Fourier and Taylor series in a certain domain.
Further, the above integrals are limited between p  x to
Open Access This article is distributed under the terms of the
p  x. This thus means we only need to consider the Creative Commons Attribution 4.0 International License (
ðp  xÞ term. In the portion to the left of limits placed, http://creativecommons.org/licenses/by/4.0/), which permits unre-
the h functions remain uniformly zero while to the right of stricted use, distribution, and reproduction in any medium, provided
you give appropriate credit to the original author(s) and the source,
the limit, we obtain functions which are a difference of the
provide a link to the Creative Commons license, and indicate if
2h functions at limit values, giving us a function one order changes were made.
lesser than required.
Thus on expanding, it looks very similar to the Taylor
series, but not quite of the same form. However, on closer
inspection, we get the same form between p to p.

2
Actually there is some subtlety regarding the definition of the
Heaviside function exactly at zero. But we’ll not need it for our
current purposes.

123

You might also like