Mod-01 Lec-01 Fundamental Theorems Connected with Zeros of Analytic Functions

nptelhrd
41 followers
1 Video View·Jan 11, 2024
Advanced Complex Analysis - Part 1 by Dr. T.E. Venkata Balaji,Department of Mathematics,IIT Madras.For more details on NPTEL visit http://nptel.ac.in
ok so
welcome to ah this course on
advanced complex analysis so
um what we intend to do is to give you
as
given this course a selection of ah
topics from advanced complex analysis
so um so of course we assume that
you already done a first course in
complex analysis basically you know
covering
the notion of an analytic function
and then
koshi's theorem and then
the idea of
taylor series a lot of series the idea
of singularities in the residue theorem
okay so ah
of course
we
we have chosen
for the topics to be presented certain
important theorems certain landmark
theorems
which are usually not
stressed upon in a first course in
complex analysis and whose proofs are
also not all that easy ok
ah but they are very
interesting theorems and they are of
very geometric nature and that is what
we will try to cover
so
so the
of course so let me start with
what we will be
doing in the first
few lectures
and that is about
trying to look at zeros of analytic
functions ok so basically
you know ah
so so we are interested in
zeros
of analytic functions
so
ah this is a broad topic for the first
few lectures all right
and of course there are what what i am
going to do is ah state some important
theorems ah connect connected with
this theme okay
so of course let me ah first of all
remind you
and
when when i say analytic function i
think of a function which is defined
on a
on a domain in the complex plane okay
so a domain is an open connected set ok
so
the fact that it is open means that
given every given point in the set there
is a small disk surrounding that point
which is contained in that set so
ah the fact that
set is open is being
is the same as
saying that the set is a union of disks
okay
ah and
of course you know we always work with
open sets because
if you want to study the properties of a
function at a point especially if you
want to take a limit
at a point then you should be able to
approach the point from all directions
and so you must have a nice disk
surrounding that point where the
function is defined so you can actually
take the limit
ok so we always study only functions at
points where
in a neighborhood of which the function
is defined okay and that is the reason
why we always study
functions defined on open sets and of
course we also study functions defined
on connected sets
because
i mean if a set is not connected then it
falls into two pieces and essentially a
function on such a set is
a different function on each piece
okay so you can reduce the study of
functions to
just studying functions on a single
piece
that is why we always study
functions defined on open connected sets
which are domains all right now
so ah
we take a
function defined on a domain in the
complex plane and we of course assume
the function takes complex values so
again
ah the the co-domain of the function is
is complex numbers and
if you remember for the from the first
course in complex analysis
there are several ways of trying to
define
when the function is analytic at a point
in the in the domain so of course
the the simplest definition is
that of
and it is the most common definition it
is that the function should be
differentiable
at not only at that point but in a small
disk surrounding that point okay
and we also use the word holomorphic
function instead of the word analytic
function that is common in the
literature
and
and we say a function is holomorphic or
analytic on the whole domain if
if it is analytic at every point of
course
the function being
analytic at a point ah can also be
described in several other ways
ah one way is the way that i told you
that it is differentiable neighborhood
of the point
ah
that is the first derivative exists in
the neighborhood of the point
the the other way of
defining a function to be analytic at a
point is
saying that the function can be
expressed as a convergent power series
centered at that point in a small disk
surrounding that point and this should
and if this happens for every point
then we could call that function an
analytic function so so basically there
is one definition of analytic function
which says that
the function is differentiable once
in the neighborhood in in a small
neighborhood of the point
there is another definition which says
that
it is represented by convergent power
series
centered at that point
and the relationship between these two
definitions is that
they are equivalent and that is the
great thing about complex valid
functions
because power series
if you would have learnt in a first
course in complex analysis is infinitely
differentiable
within
its region of convergence the region of
convergence of a power series is always
a is a disk centered
at
the point and then probably some points
of boundary may or may not be included
but within
the
disk
the power series always represents an
analytic function
and not only it is once differentiable
it is differentiable infinitely many
times so this is one of the
striking features
that differentiate real value
differentiable functions and complex
valued differentiable functions
okay so
if you if you assume a real valued
function
on on a subset of the real line c on an
interval open interval
is differentiable throughout the
interval there is no reason that
other higher derivatives exist in fact
there is no even there is no there is no
reason even that the derivative is
continuous
whereas if you assume
a complex analytic function of a complex
variable
on at a point
is ah is differentiable in neighborhood
of the point the amazing thing is that
it becomes infinitely differentiable
which means all the derivatives of all
orders exist and they are all continuous
this is the greatness
this is the amount of power that
that one time differentiability gives
you infinite differentiability in the
neighborhood okay and that's the
characteristic
i mean feature of studying analytic
functions and of course the the power
series if you take the coefficients of
the power series they're going to be
related to the
they are going to be related to the
taylor equations okay
and
these taylor coefficients can be gotten
by ah
using
cauchy integral formula ok
so so you have this notion of an
analytic function
either you define it as something that
is
locally given by a convergent power
series or
something that is a function that is
differentiable
everywhere
ok differentiable ones everywhere
and of course the usual way of checking
that a function is
is analytic is the
is by checking the so called kaushiriman
equations so
what you do is that you check the you
take the real and imaginary parts of the
function
this is how you try to check a function
is analytic
usually you take the real and imaginary
parts of the function and then
you write down the kosher remain
equations and then you check the cauchy
remain equations satisfied and
then you also
probably
check that the
first partial differentials are
continuous
and then you conclude that the function
is analytic
now
so there is a way of checking the
function is analytic using kaushiriman
equations as well
but nevertheless ah
the point is that
we are interested in
zeros of analytic functions
and
ah the first important fact that ah all
of you should have studied in a first
course in complex analysis is that the
zeros of an analytic function are
isolated ok so that is the first
important fact that means given a zero
of an analytic function
ah
there is always a small disc surrounding
that zero where there are no other zeros
ok so this is called
so if you have different zeros
they can be separated from each other by
small open disks centered at those
at those zeros
and this is
this is
what we say
this is what we say ah this is what we
mean when we say that the zeros are
isolated so the zeros of analytic
function are isolated okay
now
you see
ah so then of course that comes the
question
when you what is the problem with
looking at the zero of an analytic
function
well you take a
well you take a function which is having
a zero at a certain point which is
analytic at that point
if you take a small neighborhood you
know there is a small neighborhood where
there is no other zero because zeros are
isolated now if you invert the function
in that ah
disk
then you know the the the reciprocal of
the function is defined except for the
zero there ok and
then this gives rise to a pole
at that point and this is one example of
what is called a singularity okay so
analytic functions
there are bad there are points on the
boundary of the region where the
function is analytic
where
the reality functions are supposed to
have singularities so there is so called
singular points and again you would have
studied about singular points in the
first course in ah complex analysis
so basically one is always
one always worries about
so-called
isolated singularities because one does
not want to the the case of non-isolated
singularities is far more complicated
to
analyze so for example if you take the
function
log z
then you know
it has several branches
you have to define various branches of
the logarithm but to define a branch of
the logarithm you will have to make a
slit
in the on the plane
for example you have to slit the plane
along the negative real axis
and then you can define a branch of the
logarithm
and then the
the whole negative real axis becomes
points of singularity for this function
so
this tells you that the singularities
are not isolated because
they are they continuously lie on the
negative real axis but of course these
are not the kind of singularities we are
interested in
one always studies ah isolated
singularities and isolated singularities
basically are of three types
if you recall
ah the first one is
called the removable singularity
ah and a removable singularity is
essentially the ah is a is a nosing is
is a non-singularity okay
for example a function like sine z over
z
that is one by z times sin z
if you look at z equal to zero
if you try to directly substitute in the
function you will get zero by zero which
is which is not a defined value but of
course you know limit as z tends to zero
sin z z is one so if you define the
function to be
ah to take the value one at z equal to
zero this gives rise to an analytic
function
and
therefore the point z equal to zero is
what is called a removable similarity
for the function f of z is equal to sin
z by one by z into sine z
and
how this is reflected is of course its
reflected by ah looking at the power
series expansion
if you if you take the power series
expansion for sin z and divide that by
out by z you see that
you essentially don't get any
negative powers of z
and that tells you that essentially ah
this is a taylor series so not a lauren
series
and therefore this is not really a
singularity
and therefore and you will also see that
if you take the power series for
sine z at the origin
and divide by z and put z equal to zero
you will get one and that will tell you
that ah one should be the value that you
should define for the function to become
analytic
at the point at the origin ok so this is
what is called an ah
an isolated removable
singularity then of course come the so
called poles
of a function and the poles are well
they are they are supposed to be thought
of as zeros of the denominator ok so i
mean the simplest example is you take if
you if you are given a point ah
e z naught then you look at the function
one by z minus z naught to the power of
n where n is a positive integer and then
you know z ah z naught is a zero of set
of this function
of the denominator of this function
which is just z minus z naught to the
power of n it's a zero of order n so one
by z minus z naught to the n
has as a pole of order
ah n
at z naught ok so ah the pole is
basically a zero of the denominator
that's how you should think of it and
well
so a pole is really a singularity
it is something that you cannot
tinker with to make the function
analytic at that point okay and and
in a way a worser kind of singularity is
called an essential singularity and
that's a singularity
for example you take e power you take
exponential of 1 by z
at z equal to 0
that is an essential singularity and
both poles and essential singularities
are really the bona fide singularity the
removable singularities are actually
non-singularities
because you can always
get rid of them you can get rid of
removable singularities by redefining
the function at the point but you cannot
get rid of a pole or an essential
singularity
at a given point okay
and
of course you would have also learnt
how do you
distinguish between
a pole and an essential singularity
and they
there is a so called
laurent theorem
which is an analog of the ah or you may
even call it an extension of the taylor
theorem so the taylor theorem ah is a
theorem that if a function is analytic
at a point
then you can express it as a convergent
power series around that point
that is you can find a convergent power
series centered at that point which
point wise converges
to your given function in a good
neighborhood of the point
this is taylor's theorem and this is the
this is the theorem that actually tells
you that once differentiability
implies infinite differentiability this
is what gives you the equivalence of
the
seemingly weaker definition of
analyticity being once differentiable
and the stronger definition of
and the stronger
you know
implication that
one's differentiably implies infinitely
infinitely many times differentiable so
that's the taylor theorem but ah the
lawrence the laurent theorem is
a kind of extension of the taylor
theorem it tells you
that if you also include negative powers
then you can get a series involving also
negative powers
ah in a deleted neighborhood of the
point
and that is called the laurent series
and ah
for all you know the laurent series may
have negative powers ah of arbitrary
order ok so
ah you know so if you if you look at
if you if if z equal to
z naught is an ah well isolated
singularity
of f of z
ah
of course i am assuming f is an analytic
function and z naught is a singular
point it is an isolated singularity ah
then
then ah
we have a nor we have
we have a lore expansion
so
well
f of z
is equal to well let me write like this
a naught
plus a one
z minus z naught
plus a two
z minus z naught squared
and so on this is the this is what is
called the
the analytic part of the loraw expansion
and then you get the negative powers you
get a minus one by z minus z naught
then you get a minus two
by
so this is a subscript minus one its not
a minus one ok
and this is a subscript minus two z
minus z naught squared
and so on
so ah
this is called a neuron series centered
at z naught and the function converges
to this this equality means that this
normal series converges if you plug in a
value of z
in in a small disk surrounding z naught
then this series converges to a value
which is equal to the function value
at the at that point
and ah
and of course the
the the point should not be z naught
because you cannot substitute z naught
here because you will be dividing by
zero all these negative terms and the
fact and the fact that if z equal to z
naught is a removable singularity then
all these negative coefficients will be
zero so your laura expansion will
actually be a taylor expansion that is
exactly what happens when you look at
one by z times sin z for example at z
equal to zero
ok and ok so
this is so called long expansion and
the point is that
among
the
important singularities namely the poles
and the
and the essential similarities
you can also distinguish the type of
singularity by looking at the
at the laurent expansion if you get
infinitely many
of these negative powers
okay then
the singularity is an
isolated essential singularity
ah for example exponential of one by z
at z equal to zero
and if you get only finitely many of
these negative terms then it is a pole
ok
ah and the order of the pole will be
equal to
ah the minus of the
largest negative subscript you get here
okay
ah so that's another that's one way of
you know trying to
distinguish between a pole and an
essential singularity there is one more
way of distinguishing between a poor and
essential singularity and that is by
taking limits
ok
if you take the limit of the function as
the point tends to the singularity
and if the limit
exists and is equal to infinity
from all directions so of course this
means that you have to make sense of
what ah limit equal to infinity means i
mean the limit of a complex quantity
you say it is equal to infinity if the
modulus of the quantity becomes
arbitrarily large
ok no matter how you approach the
limiting point so ah you know if the
singularity z naught is such that as z
tends to z naught in no matter in
whatever direction
the mod
of f of z
goes to infinity
then you say ah
the limit of f of z as z tends to z
naught is infinity and this is a
situation exactly when z naught is a
pole
okay and if ah
there could be and what happens in the
case of an essential similarity the
limit will not exist
in the sense that you might get limits
differently limits as you approach in
different directions for example you can
take exponential of one by z and try to
calculate the limit from by approaching
the point z equal to zero from the
positive axis from the negative axis
uh real axis positive real axis you will
see you'll get different values the fact
the fact that you get different values
from different directions tells you the
limit does not
and that's precisely
that's precisely the condition that
tells you that
it is an essential similarity okay so
so the most important thing about
singularities is what is called the
the residue of the ah the residue of the
function
uh adds at an isolated at an isolated
singularity and that is supposed to be
the value of this
this coefficient a sub minus one which
is the coefficient of
one by z minus z naught ok so you know
residue
of
f of z
at
z equal to z naught
is
a minus one
ok
and this is a very very important
value for the function
because
the
because it is connected with the residue
theorem
ok ah which tells you that this is what
you will
this is what you will get if you
if you try to integrate if you integrate
the function
over a curve surrounding or sim or a
closed curve that surrounds this point a
simple for example a circle
a circular arc a
circle that surrounds this point if you
integrate the function
what you will get is
2 pi i times
ah 2 pi times this
this quantity so for example you know
you know you can take a very very simple
example take g of z to be ah one by z
minus z naught
i mean this is the simplest thing you
can think of okay and
you
well here is z naught on the complex
plane and then if you want draw a circle
ah look at the circle mod z minus z
naught
is equal to
rho
there is a circle
and
obviously this is an analytic function
except for the point z equal to z naught
everywhere else the denominator never
vanishes z naught is zero of order one
so it is a simple pole okay a pole of
order one is called a simple pole
and if it is order is greater than one
it is called a multiple pole and well if
you try to integrate if you calculate
one by two pi i
integral over
if i call this circle as gamma
ah
f
of g z
i'll end up with ah i'll end up with one
ok so in fact you know if i if you want
i can even put a lambda here
where lambda is
any complex number
okay and if i integrate it what i will
end up with is well if you
if you would have done this several
times
so if you want to integrate over a
contour
the method is that you first
parameterize that contour and then you
make a change of variable mind you
whenever you integrate something
when you integrate a function or a
contour you must understand that the
variable lies on the contour
ok so that means that you should write
an equation for the contour and that is
called
parameterization of the contour so the
parameterization of this contour is z
equal to e z naught plus rho e to the i
theta
where theta
varies from zero to two pi
so
this integral becomes well if you write
it down
perhaps you already done it
you certainly have done it but let me
just
recall
quickly
you you will just get lambda so you will
get 1 by 2 pi i
integral theta from 0 to 2 pi
i am going to get
g of ah
z is z naught
plus rho e to the i theta
ah
of course i forgotten to write d z there
which is the
variable of integration
ok so i will get d
t
uh z naught plus rho e to the i theta
and this this will turn out to be one by
two pi
integral zero to two pi
what you will get here is
one by
ah z minus z naught if i substitute this
i get rho e to the i theta
and if i differentiate this
i will get
ah i
uh
rho e to the i theta d theta
ok
and
what will i end up with i will end up
with when ah
well
my root of the i theta cancels my i
cancels so i get one by two pi i
integral zero to two pi
i think i forgot a lambda there so there
is a lambda on top
so i will get lambda p theta and that is
just lambda
ok
so
ah so the moral of the story is that you
see if i look at this function
and integrate it over a small ah
over the small circle surrounding this
ah
this
this point which is a simple pole i pick
up
this this this coefficient and you know
actually
about this point if you try to
write the laurel expansion this is a lot
of expansion
the lauren expansion for lambda by z
minus z naught is lambda by z minus z
ok
it is already the loral expansion and a
minus one is
the coefficient of one by z minus z
naught and that is lambda and that is
what shows up when you calculate 1 by 2
pi and that is a residue for the
function so this is this is a
simplest illustration of
the philosophy behind the residue
theorem the residue theorem says that
if you integrate
over
you know
the point which is isolated singularity
and
assume that there are no other
singularities
then what you will get is the residue
and of the you will get
i mean 1 by 2 pi so let me write that so
to be more precise you have the residue
theorem
which is the
ah
residue theorem is a starting point
fatty is for ah
our discussion so you know
so basically you have
let us assume that you have
a nice contour
like this
and you have a function
f defined on this
ah on this
on a on a domain which contains this
contour and the interior of the contour
and assume that you know there are there
are
well
several isolated singularities that one
z two
and so on
zn
if you integrate the function if you
write 1 by 2 pi i
integral over
this contour
of
f z d z
what you will get is ah
some
summation
ah i equal to one to n
residue
of
f of z and z r
this is the residue theorem
so what i have done here is i have taken
ah
i have simply taken
the function to have only one
singularity
isolated singularity and i do this
integration and what i end up is the
residue at that point but if you have
several and of course you should assume
that there are no
there are no singularities on the on the
contour over which you are integrating
okay so you know
the assumption for this is that the
function is analytic in the interior
and also on the boundary
which means that
to say the function is analytic on the
boundary means actually it means that it
is analytic in a small disk surrounding
every point on the boundary which means
it is actually analytic in an
in a bigger open set
a bigger domain which actually contains
this boundary and the interior okay so
then this is the so-called residual
theorem and
in the simplest case it reduces to this
and you can also see that you know
if you
if you take this function
and you
instead of taking lambda by z minus z
naught suppose i took this power series
i mean if i not power series if i take
this lawrence series and i integrate it
okay around the contour
like this
ah then of course
the first thing is that the integral of
this whole series
is the same as integrating term by term
that is you can integrate term by term
and then
take the resulting series and this is
pau this is correct because
you can interchange integration and
summation
provided the
series of functions converges uniformly
ok and it is a theorem
that if you take a long series then
within
in the reason why the long series is
defined if you take a close disk in that
region then the long series will
converge uniformly
ok and of course whenever i say long
series should be it should be a deleted
neighborhood
you should not include the point of
course because you cannot substitute the
point
because you'll be dividing by zero for
the negative terms but but of course
there is a similar theorem for power
series which says that whenever you have
a power series which is converging in a
in a disk
then you take any
closed disk inside that disk the power
series will converge there
in fact absolutely and uniformly okay so
so because of this uniform convergence
the integral if i calculate this
integral for this function i can
actually integrate term by term and you
know if i integrate term by term from
here onwards each term will give me zero
because its cautious theorem cautious
theorem says that if you integrate a
analytic function
uh or a simply closed curve
there is
the i mean the the integral is going to
vanish
you are not going to get anything ok and
so the integral of all these terms will
go away okay
the integral of this term will give you
a minus 1 of course
okay and the integrals of all these
terms will also go away because
one by e z minus z naught to the power
of ah for example one by z minus z
naught the whole square has an
anti-derivative which is just one by
minus one by z minus z naught to the
power of one
okay so all these negative terms from
power two onwards they all have
anti-derivatives and
it's a it's a it's a version of the
fundamental theorem of calculus
that whenever a function has an
anti-derivative then ah the the the the
integral is just the anti-derivative
evaluated at the final point minus the
anti-derivative evaluated at the at the
initial point but then this is a closed
curve the final point is same as the
initial point therefore you get zero so
if you integrate this
term by term
the only thing you will pick up is a
minus one
and that is the proof of the residual
theorem if you had a single
ah if you had a single
singularity okay and then if you have
several singularities
this follows because
of cauchy's theorem because what
cauchy's theorem will tell you is that
cauchy's theorem tells you basically
that if you take an analytic function
and integrate
it over a simple closed curve at the
integral of zero okay but this is a so
called simply connected version of
koshis theorem which is over a simple
closed curve
ah where your dom which means
the the region inside the curve has no
holes
okay but then there is a new there is a
different version of questions theorem
which says that you know if your domain
is has an outer curve like this and you
have inner curves
you have inner curves
and of course you know in all these
issues
the orientation of the curve is very
important we always take the curves to
be
oriented in the anticlockwise sense and
that is called the positive orientation
okay and if you change the orientation
the sign of the integral will change
that is how it goes
and if you if you apply cautious theorem
to the region
ah
which is
this the interior of this curve
and the exterior of all these little
curves that the function is of course
analytic and therefore you will get that
the integral is zero but that will
amount to say that the integral over the
the outer region the outer curve is the
sum of the integrals of the inner curves
okay and
but then sum of each in but each
integral ah will give you the residue at
that point as i have explained here
and therefore you get the residue
theorem
basically you get the residual theorem
from
an application of cauchy theorem and
literally
this this kind of argument okay
fine so ah
so you have the residue theorem now
the
you see the the
ah so let me
having told you
so far let me also tell you ah what kind
of theorems
we are going to
prove okay
i i think i i do not know how many
lectures it may take but
probably a few lectures so you see the
kind of theorems we want to prove
are actually theorems about zeros of
analytic functions and well
so ah
yeah so
a glimpse
of the
we would like to prove
so
probably some of you who have done a
little bit of further reading beyond the
first course might have seen proofs of
these theorems but
this is where i would like to start the
course ah so the the first theorem is so
called
argument principle
ok
and the
this is so called argument principle
and
ah
the
the
and what is this argument principle so
so let me state the theorem that i
that i want to prove one is the argument
principle probably its just
some kind of
corollary to the residual theorem if you
look at the logarithmic integral
if you remember it from the first course
but anyway i will recall it
the the second thing is using the
argument principle or even otherwise you
can prove the so called roches theorem
ok
ah
then we would like to study
we would like to prove herbert's theorem
and of course
and thereafter one would like to
study
you know
the open mapping theorem
to prove the open mapping theorem
ah
and
of course also one would like to prove
the inverse function theorem
ok
so the ah
these are the
i mean the first set of theorems we
would like to prove okay probably you
would have seen the first and the second
maybe
but anyway its
they are the starting point so i will
make it a point to recall them ok so let
me explain
let me explain what these what these
theorems are
so
the first one is argument principle
so what it
ah so what does so what does it say
ah so i briefly describe
what the statements of these theorems
are
and you will see that they are actually
connected with i mean they are the right
theorems that
that will come under this topic
trying to study zeros of analytic
functions ok
so
the argument principle
so the argument principle is well
you know ah
1 by 2 pi i
integral over
a simply closed curve
so when i say simple closed curve ah
let me explain what that means first of
all a curve is said to be closed if its
initial point is the same as a terminal
point ok of course by curve generally we
mean the image of
an interval
a closed interval if you want
the closed interval 0 1 on the real line
a continuous image of that on the
complex plane is called a curve for
example
the circle is a curve because it is the
image of the interval 0 2 pi
under this function
theta going to
if you want z naught plus rho e to the i
theta ok so
it is a continuous image of the interval
and its a closed curve if ah the initial
point is the same as the final point the
fact that there is an initial point and
there is a final point tells you that
the curve is already oriented
okay that means there is a direction for
the curve and that is that direction is
given by the direction of
the that that is given by the the the
direction of increase of the
the parameter the variable that is used
to write the equation of the curve okay
and
ah and when i say simple curve it means
that the curve is doesn't it cross
itself doesn't intersect itself so it is
not something like a figure eight or
more complicated curves that cross
themselves ah
one segment of the curve ah twists and
turns and come comes back and
hits itself at some point again crosses
itself ok there are no such self
crossings so such a curve is called
simple curve and
ah since you are going to do
ah since you are going to do integration
ok the curves that we will always deal
with will be piecewise smooth that means
the curves ah
if you write down the parameters
parameterization for the curve okay
then the parameterization will always
come
will be defined over some interval
and the fact that
you can divide this interval into
sub intervals
in each of which the function you write
down
is actually differentiable
it is differentiable and continuous
ok so this is what is called a piecewise
smooth curve so for example here the
function theta going to z naught plus
rho e to the i theta where theta lies
from zero to two pi is of course ah
continuously differentiable function of
theta which is the parameter okay but in
more but more generally uh
curve
need not be
given by single parameterization it
could just break down into several
pieces and each piece may have a
different parameterization
ok one piece may be say part of a
circle another piece may be part of a
parabola the third piece may be part of
the line but it does not matter the
point is piecewise it has to be it has
to be smooth
so so whenever i say simple closed
contour when i say whenever i say
contour its always a something that is
ah piecewise smooth ok so the argument
principle basically tells you that if
you are looking at
a function which is
a function defined on a domain like this
okay
with the property that
this
the function is analytic on the
on this
on on a domain which contains this whole
region
okay
except at
finitely many points
which which line the interior
which are ah
only poles
okay
you you assume that they are only poles
okay and the function should not have
the function should not have any zeros
on the on the on the boundary of the
contour okay then
the one by two pi i integral over gamma
of d log f z is called the logarithmic
integral
ok which is ah
will give you
the number of
zeros minus the number of poles
inside the
inside region
ok so this is the this is basically the
argument principle and and of course
d log fz of course means
d log
f z means
f dash of z by f set
what you must understand is that because
f is analytic wherever f is analytic f
dash is also analytic
because as i told you a function that is
analytic is infinitely differentiable so
and this is a quotient of analytic
functions it will be analytic there the
only problem is the denominator might
vanish
so wherever you have zeros of f
f dash by f the logarithmic the so
called logarithmic derivative of f
will have a pole
wherever f has a zero
and
of course
if f
has a pole
then f dash have a pole
ok so the only poles for this function
we assume the only poles for this
function are some zeros of f inside zero
should not lie on the boundary
and ah some poles of f
inside and they also ah should not lie
on the boundary so the boundary should
be four free from both zeros and poles
and there are only finitely many zeros
and poles inside
inside the
inside your domain okay
and
so this is the argument principle so
computing the log if you integrate the
logarithmic derivative
the argument principle tells you that
you get the difference between
the ah number of zeros and number of
codes so that is the that is the so
called argument principle okay and then
let me quickly tell you
about
what these other
theorems have to say
so ah
so i am
i am just now giving a an overview of
results but then we will
go into them more deeply so what is
rosh's theorem so rosh's theorem
is ah so as i told you ah
this this whole
exercise is to
somehow uh
study zeros of analytic functions okay
and
basically for example you want to count
the number of zeros in a insider inside
a in a region which is bounded
by a lets say a curve a simple closed
curve
so ah the point is that
the point is that you do not get ah
in general a zero of f
could be a will of course be a pole for
f dash by f so you cannot avoid
considering poles also ok so that is the
reason the argument principle gives you
the zeros and the poles also so in
particular if there are no poles
then you will be counting the number of
zeros and of course when you count
number of zeros mind you every zero has
to be counted with multiplicity so for
example you know if you take the
function lambda by z minus z naught this
has zero of order one at z naught
ah so the number of zeros will be one
if you take a ah
you have simple close curve enclosings
are not
if it is not going to enclose that not
the number of zeros will be zero
ok but if i replace this by lambda by s
minus n naught to the power of m
then the number of zeros will
be m
though physically there is only one zero
at z naught but its order is m
so its also the order of
vanishing
okay it is the number of times the
factor
z minus z naught z minus zero
appears
so you whenever you say zeros
ah or poles you have to count them with
multiplicity
zeros also have to be counted with
multiple multiplicities for example if i
take the function z minus z naught to
the power of m
where m is positive
then z naught is a zero of that function
but
its it should be counted as m zeros
so zeros and poles have to be counted
with multiplicities and then this
formula holds so this is a kind of
counting principle okay
ah
that's one thing then
rosh's theorem is something more
the philosophy of rosh's theorem is that
you know you take a
you take an analytic function
in a domain
and suppose you are interested in the 0s
inside
inside the region the domain that is
you know
that you get
in the interior of a simple closed curve
ok
then rogers theorem says that you know
if you
if you perturb the analytic function
little bit
okay
even after perturbations the number of
zeros will not change
ok that is if you
if you add to the analytic function
another analytic function
which is which is small enough
that means you add to the analytic
function as smaller analytic function
of course you know there is nothing
called smaller or bigger and complex
numbers because complex numbers are not
ordered but then whenever we say smaller
or bigger we always refer to the modulus
so you know what roger's theorem says is
that you take a function f of z which is
analytic ah in a in say a
in in say a bounded region
surrounded by simply closed curve
ah then the number of zeros
will be the same
for f of z and f of z plus g z
where g z is a smaller function
smaller on the boundary
ok so that is rochester
and you think of adding g to f as a
small perturbation of f
ok so so i i just write it in words
the number of
zeros of zeros
is not affected
inside in inside
a
simple
closed contour
is
invariant
is invariant means it does not change
is invariant
under
small
perturbations
so you take the analytic function
and add to it a smaller analytic
function a function that is smaller than
this function on the boundary
on the boundary control
then even after adding it the number of
zeros is not going to change
okay
so ah the the addition of another
analytic function which is dominated by
the given analytic function on the
boundary is called perturbation if you
want
okay and its a small perturbation
because what you are adding
ah in modulus is strictly less than the
modulus the given function on the
boundary ok so this is this is
rochester's theorem right
so
ah one version of roche's theorem will
tell you that suppose you want you have
two analytic functions f and g
ok
how can you conclude that they have the
same number of zeros
ok
so the answer to that is ah you
calculate you know if you if you
calculate mod of f plus g
triangle inequality will always give you
mod of f plus z is less than or equal to
mod f plus mod g okay now the question
is on the boundary if you get strict
inequality if you get mod of f plus g is
strictly less than mod f plus mod g on
the boundary
then both f and g will have the same
number of zeros inside
that is another that is another avatar
of this rosh's theorem so it helps you
to com also to compare
the it tells you that two
analytic functions will have same number
of zeros if ah the sum of their moduli
dot is strictly greater than dominates
ah
the modulus of their sum on the boundary
okay that's another author of this okay
so it helps you to compare number of
zeros
then the third one is
hurwitz theorem
so this
this heritage theorem is again you know
it's it's again a very beautiful theorem
what it says is that you know if you
have if you have a
if you have a sequence of
analytic functions
which is converging to a given function
okay in a domain
and assume that the convergence ah
is going to be uniform on every closed
disk in the domain so this is called
uniform convergence on compact subsets
okay
the other word that is used in the
literature is called normal convergence
okay so if you have normal convergence
ok
which means uniform conversions of
compact subsets for example the
convergence is uniform on every closed
disk
in your in your domain
so if this is if this is a normal
convergence
and
and if
ah f has
a zero
of order n
at
z naught
then
what happens is i i'll draw a diagram
so e z naught is
is a point where the limit function f of
z has zero okay now
some fundamental complex analysis will
tell you that because the conversion is
uniform and since each of these
functions is already analytic f will
also turn out to be analytic
ok
ah that's again an exercise
that you can easily try to do
and in fact the derivatives of all these
will also converge to the derivatives
of f ok and all this is just because of
this normal convergence uniform
convergence over compact subsets
and
and of course in the k in in that case
the integrals will also converge because
the moment you have uniform convergence
integrals derivatives everything will
behave well with respect to limits ok so
ah so suppose the limit so the limit
function will be analytic
and if it has a zero of order n at the
point z naught what will happen is you
can find a small enough
disk surrounding z naught
ah
small enough radius such that
beyond a certain stage all the fns
they will all if you take each fn
it will have n zeros inside this
each fn will have n zeros
with multiplicity which means that some
of these zeros may be multiple zeros and
the beautiful thing is as you if you
plot those n zeros
ok
ah
so i i
i was so you know if you plot those n
those capital n zeros and ah if you make
this small n become larger and larger
then these
these various zeros will slowly come
they will coalesce they will all tend to
the points that not
ok so what it tells you is that
when you take a
nice limit of analytic functions
then the zero of the limit
comes from zeros
is a limit of zeros of
the same
ah number of zeros
of the functions in the sequence
ok so
if f has a 0 of order n it is z naught
then
then you know
then let me write it out
somewhere here
so
then
there exists a rho greater than 0 such
that
for large
n
f n has
n zeros
in mod
z minus z naught
less than rho
ah
which
con which which
which converge
to z naught
as n tends to infinity
ok so what which theorem says is that
the zero of the limit comes
by you know you take 0 of beyond a
certain state it is the zeros of the
functions that are giving the limit
zeros of the functions that you are
taking limit of its those zeros that
slowly you know coalesce together and
give you the zero of your
function
okay so this is ah this is hurwitz
theorem
and
then
i quickly tell you what the open mapping
and the inverse function theorems are
the open mapping theorem is a very
beautiful theorem it tells you that
if you take
an analytic function
ah
and you take a point where the function
is nonzero
i mean the derivative of the function is
not zero
then
there is a neighborhood
of that point
where the function is an open mapping
which means that it will map open sets
to open sets and this is a very deep
result ok because it is very rarely
need to polish that you get open maps
ok and
for example a bijective continuous map
need not be a topological isomorphism
but if it is bijective continuous and
open then it is an homium offset because
openness of the map along with the
injectivity will tell you that the
inverse map is continuous so you know so
what it tells you is that an open map ah
is
as good as an isomorphism except that
you need to know that it is injective if
you know its injective then its an sum
of sum ok so you can invert the map and
all this is true analytically complex
analytically also so the only condition
is that you know the derivative should
not vanish at a point and then in
neighborhood of the point
uh everything is beautiful you it's a
local isomorphs okay that's essentially
the open mapping theorem and
ah let me quickly tell you about the
ah
the inverse function theorem the inverse
function theorem is that
again
in a sense ah
you can you can think of it as another
variant of the open mapping theorem what
it says is that whenever the derivative
is non zero ok at a point then there is
a small neighborhood where you can
invert the map ok you and and the the
inverse function can be written again
using a cauchy integral ok there is a
there is an integral formula to write
out the inverse so you can invert the
function there is an explicit formula ok
so
so the so this is the so i am not i am
not
writing more details we will go in we
will go into them in the succeeding
lectures
so the point about these two theorems is
that
you must be essentially looking at
ah
well there is a more
general version of open mapping theorem
will tell you that
all you do not even need the derivative
to be non-vanishing essentially you need
a non-constant analytic function ok so
ah
you need a
you need a non constant analytic
function and it will always map
ah open sets to open sets ok and in
particular if the
if the if the derivative is non zero
then in a neighborhood
of that point where the derivative is
not zero the function is actually a
holomorphic isomorphs an analytic
isomorphism which means it is an
isomorphism
its injective
ah on to its image which is open and if
you take the inverse function that is
also holomorphic that is also analytic
and that is given by a nice formula
integral formula that is what
ah the inverse function say theorem says
okay now all these are somehow connected
with zeros of analytic functions and
they can all be derived starting from
the argument principle which essentially
is
i should say
the residual theorem
applied to not to f
but its residue theorem apply applied to
the logarithmic derivative of f so the
root for everything is the residue
theorem ok
so we will do this in the forthcoming
lectures right so and let me stop here
you
