Lecture 01

IIT Kanpur July 201
20 followers
Jul 19, 2024
[Music]
[Music]
welcome to the first
lecture uh for the course of forer
analysis this is going to be an
introductory lecture and where we will
find
the origin of of the four year
series and uh what are the we'll discuss
about PR
recites we need for this
course
uh so
first this is Joseph fer so
who had initiated study of for year
series and which has profound impact in
mathematics in every sphere of
mathematics so this is the picture of
fuer fuer was born in
1768 in
oxer
France he was ninth child and was orphan
at the age of
10 the French Revolution prevented
forer's desire to enter the Benedictine
teaching order which was basic basically
reserved for elite people and fer was a
mathematician and physicist and is
renowned for his groundbreaking
contribution to the study of heat
transfer and mathematical
analysis his famous work on the foure
series revolutionized our understanding
of periodic functions and paved the way
for numerous application in Science and
Mathematics as well as in engineering
so let's see first all of us we know
about the heat equation so the heat
equation is uh given by D by DT some
constant time d by DT d s u by D x² + d
s u by D y² in the plane for some
constant
C so when the heat exchange stops
reaching the thermal equilibrium that
means d by DT is equal to zero then this
equation the steady it's called the
steady state
equation and is obviously given by d s u
by D x² + d s u by D y² is equal
to0 now if we are looking
the heat equation on the unit disk that
is this is 0
0 and this radius is one the unit
disk uh this is lesser equal to 1 if I
take the d is this is less or equal to
1 then this
is the our unit
disk which we can write it in terms of
the polar coordinate so now if in the
unit disk if we are taking any point
here then this is going to be rep can be
represented by the radius which is r and
then this is
Theta so that essentially is now R can
vary from 0 to 1 if we are taking the
open unit disk and which means the
boundary is not
there and then the laian is defined as
Dela is equal to d s u by D x² + d s u
by D
y² in the polar coordinates the LA is
going to look like
d s uu by D r² + 1 by r d u by d r + 1
by r² D u² by Del Theta square and this
you can easily derive by making the x
coordinate in the cartisian x coordinate
as R cos Theta and y coordinate as R sin
Theta then apply the rule of the
differentiation and but easily with
little bit of Pence you can compute that
leian is going to look like this so the
steady state heat equation means Delta U
the laian of U is equal to
zero okay so now that's what we want to
now solve this problem so now when Delta
U is equal to Z so in polar coordinates
we can see that d s u by D r² + 1 by r d
u by d r is equal to - 1 by r² d s u by
Del Theta s if you multiply R square on
both the side what you are going to get
is this
equation so now let's make an attempt to
solve this equation by using the
separation of variable so let's assume
that the
solution of this can be separated in R
variable as well as in Theta variable so
let us assume that if our solution looks
like U Theta = to f r in G Theta
then or how what is going to be the
value of f and g okay so now if I take
the derivative with respect to our
variable then we are going to get F
Prime of R because now f is a function
of r so this is f Prime of R into G
Theta and if we take the second
derivative with respect to R we'll get F
Prime of r g
Theta now if we similar if we take the
derivative with respect to Theta then f
is not going to get affected and hence
uh we will get the derivative with
respect to Theta will be transported to
G and uh to distinguish between the
derivative with respect to R and Theta I
am uh writing the symbol
g double dot Theta means del d s g by D
Theta
squ okay
so therefore we
have by this equation if we put it then
we get that this is
r² f Prime r + r f Prime of r divided by
F this is equal to - G
derivative with respect to Theta by G
Theta uh look at the both the left and
right hand side now the two sides are of
different variable left hand side is
independent of theta and the right hand
side is independent of
R therefore this ratio is going to be a
constant so what we get from this is
equal to
Lambda means
g double derivative with respect to
theta plus Lambda of G Theta is equal to
0 and in if this is equal to Lambda
r² left hand side is equal to Lambda
then we get r² F Prime of r + r f Prime
of R minus Lambda f r is equal
to0 since we are
in the disk and this Theta variable is 0
to 2 pi and then g g is a g is going to
as a function of theta is going to
repeat after each 2 pi
interval so we call that to be a 2 pi
periodic function in this course we'll
discuss uh a lot more about the periodic
functions at a later stage in this
introductory lecture I mean so I presume
that uh you understand that 2 pi
periodic means this is going to repeat
after after 2
pi
so and now if we take Lambda to be less
or equal to zero then one can easily see
that what we are going to get is Trivial
solution therefore we assume that Lambda
lamb is greater than zero now that means
Lambda and it has to be an integer so
Lambda equal to m² for Su M belongs to
zed therefore what we got is that if we
solve this equation then what we got is
G Theta equal to AA cos M Theta + blda
sin M Theta for some scalar AA and blda
now we know theil's identity e^ I Theta
isal to cos Theta + I sin Theta then we
have if we plug in we can write G Theta
equal to a e^ im Theta + B e^ minus I IM
Theta for some constant A and
B now Lambda is equal to m² so we need
to look at what what is going to be the
form of the solution capital F the
function so we have this equation r² F
Prime r + r f Prime of R minus Lambda f
r is equal to
0 okay now if m is not equal to
zero then F of R is equal to r^
M and F of r = r ^ minus r of M so this
is if I plug in R the power M you have M
into M -1 R the
power nus 2 you multiply this you get a
to the^ M here you are having M and so
this is equal to
0 now if m is equal to Z then obviously
F of r equal to 1 this is going to give
us a solution as well as it can give a
solution that F of r is equal to log R
if you take the double derivative then
easily one can verify that this
satisfies the equation here that is easy
to
see now if m is greater than zero then R
the^ minus of
M goes to Infinity as R goes to zero
because this is going to be 1 by r power
M so when R goes to Z this is going to
be infinity and if m is less than Z then
R to the power M goes to
Infinity as R goes to zero so that
that's why if m is and if m is equal to
zero then log R also goes to Infinity as
R goes to zero and remember that we are
looking for bounded solution therefore
for each M belongs to
Jed we get that we can discard this so
we get u m r Theta = to r^ mod M E power
I IM
Theta now what is the claim of
fua the general solution what we get is
U Theta for each M we are getting a
solution and this is a linear equation
so we get the general solution we can
write that M = minus infinity to
Infinity a m r the^ m e to the^ I IM
Theta and where am is a scalar now you
can see that uh if for each fixed R if
am is a bounded if
a is less or equal to some k for all M
belongs to zed then the series converges
and that is what is going to be the
general solution of uh the heat equation
steady state heat equation if this now
remember that we had assumed that if our
solution looks like f of r g Theta then
this will be of this form if this gives
all the solution to Delta U equal to Z
that means what we are saying that if we
have solution of the heat equation
steady state equation heat equation then
it must look like Ur Theta is equal to f
r g Theta then in this case if I put in
the boundary where Ral to 1 is the
boundary of the disk then U of 1 Theta
is equal to a m e^ i Theta which is a
function of theta no more dependent on
the radial variable
so now this is a function which is
defined on the
circle so so given any reasonable F on 0
to 2 pi and which is f of 0 is equal to
F of 2 pi that means we
are talking that f is a 2 pi periodic
function and now the question comes that
can we find a m such that
F of theta = to a m e^ IM
Theta so this is what
fer claim that uh if we have a function
on 0 to 2
pi
then which is
periodic then this F of
theta can be expressed as M = to minus
infinity to Infinity a m e to power I IM
Theta of course now this raises many
questions first of
all that given a
f what are these am can we determine
this
am that is the first question
which
forer
answered like if for the time being if
you allow me to do certain calculation
in without caring about mathematical
rigor then you can see that 0 to 2 pi F
Theta D Theta this is equal to suppose
this expression holds 0 to 2 pi
summation / m a m e power i m th
D Theta now if I am allowed without
bothering that whether what is the
condition that we can interchange the
sum and the
integration if we just formally do this
0 to 2 pi then this am is independent
of independent of M so we get e the
power i m Theta D Theta now this
integral we know that this is equal to 2
piun if m is equal to 0 and is equal to
Z if m is not equal to Z therefore we
get this as 2 pi of
a0 now in the same way if we write down
0
to 2
pi F of theta e to the^ power minus of i
n Theta D Theta then this is equal to
without caring about how to change the
summation and the integration we just if
we do it formally then this is going to
be 0 to 2
pi E power i m minus of n Theta D Theta
which is this integral is going to surve
AR only when m is equal to n now
remember that here we are taking the n a
fixed
n so this is equal to 2 pi of a n so
without caring about uh the correctness
of all this
mathematical uh uh ideas we have
employed I mean if we forget if you
grant me that I have all the freedom to
Interchange the sum and the integral
then we know that giving a f how to
determine a but all these steps need to
be justified mathematically why all
these steps are valid
steps then the second question obviously
comes that this is an infinite series
now what is the guarantee that this
infinite series series will
converge then the next one would ask
even if it
converges suppose I have a which are
denoted by this
formula and suppose this infinite series
converges then what is the guarantee
that this infinite series is going to
converse to the original function f what
we have started off
so these questions are very deep and in
the course we are going to address all
these questions one after another so
before getting into more rigorous
mathematics so let us first recall our
uh what we have learned
in the topic of infinite series and we
will be needing here we are talking
about the integration so we will be
needing
about
certain properties of remen integration
so let's first review
quickly what we have learned for
infinite
series that's the first
one okay so by infinite series what is
it mean because we know that we can sum
we have an algorithm where we can add
finitely many terms but when it comes to
infinite at then we have to be very very
careful so what we do is that we take
the partial sum SN which is n from 1 to
n a n now this becomes a sequence of
real numbers if I am taking a n in
R and if a
n are complex number then SN is also a
complex
number
okay uh we say that the
series this series converges to
L if the sequence SN converges to L now
the sequence
is sum partial sum of a an which are
real or complex number so we know a
sequence of real or complex number what
is the meaning of convergence of a
sequence so if this partial sum this
sequence converges to L then we say that
the series we have started of converges
to
L uh so what is the meaning of the
convergence that means as usual just
recall that for a given Epsilon there
exists a n subot such that that mod of
SN minus of L is less than Epsilon for
all n greater or equal to n subn
so now to write that formally that the
series a and converges if and only if
for Epsilon greater than Z there exist
an subot such that the tail of the
series N = to K to m a n is less than
silon for all M greater or equal to n
greater or equal to N Sub Z as you can
see that if I am taking s minus of
s if s SN is a Cy sequence
then this is less than Epsilon and that
is nothing but what we are writing it
this guy is n = to K to M of
a
n so every convergence sequence is a Cy
sequence so therefore what we have got
is this that is an if and only if
condition if the partial sum is a Cy
sequence then it converges because R is
r or C they're
complete now if a n
converges if the series converges then
we know that the nth term they go to
zero just by this
because this SN - SN minus of 1 is equal
to a n then from the we can see that the
N term converges to zero and this is a
necessary condition of a series to be
convergent this is far far away from
being a sufficient condition that all of
us we know that n = 1 to Infinity
summation over 1 / n it diverges
okay so so far so good now to test the
convergence
of the series which we are quite
familiar with and uh now
if most of the test what we have learned
they are valid for if our ANS are non-
negative one of uh
very favorite I'm and most useful among
all the test
are is the comparison
test that means we have
a series which converges and that series
consists of non- negative terms
BN now we have a
n which are also non-
negative
and for each n this is less or equal to
some constant time BN and this constant
is independent of n so this is true for
all
n and if BN converges then a n converges
that means if I have the series of non-
negative terms and which is
dominated by
a BN which is a part of the converg
series then our series is convergent and
this is this is uh a very very uh useful
test
to figure out whether the series is a
convergent series is or not then uh as
an application of the comparison test
that is some easier way to deal with is
that you can find that root test that if
uh
if mod of a n the^ 1 / n is equal to l
now if L is less than one then the
series converges L is greater than one
then the series diverges and if L is
equal to 1 we can't conclude
anything and uh then the ratio test as
we're familiar with again for the non-
negative uh a if limit of a n + 1 by a n
is equal to l
then and L less than one it converges L
bigger than one it is diverges and L
equal to one we can't conclude
anything so I think this is
uh this is a good point to stop with uh
then we will see that what are the other
kinds of test in which we can conclude
whether the a series converges or not
thank you
[Music]
[Music]
[Music]
