Open in TheBrain

21 - Complex integration

00:00

welcome to the first lecture in our

00:02

fifth week of the course analysis of a

00:05

complex kind today we'll learn about

00:07

integration in the complex plane but to

00:13

begin with let's review integration in

00:15

are given a function f that's defined on

00:19

some interval a b and is real valued and

00:22

continuous we define the integral from A

00:25

to B of F of T DT as the limit of these

00:31

left-hand sums so we take f at a point

00:40

TJ multiplied it by the next point TJ

00:43

plus 1 minus TJ and we add all these

00:46

things up what are these TJ's well we

00:50

take the integral from A to B and be

00:53

divided up into these little chunks so a

00:55

is equal to t0 the next little point is

00:58

t1 then t2 and so forth up to TN and NS

01:01

on a large number and we're going to let

01:02

that go to infinity later on so right

01:05

now we're dividing up this interval

01:06

duties and chunks and we evaluate F at

01:10

the left hand point of each of these

01:13

integrals so for example f of T 0 is

01:15

this value right here we take that value

01:20

and multiply it by t1 minus t0 what is

01:23

t1 minus t0 that's this length right

01:26

here so this length is f of t0 t1 minus

01:30

t0 is that length so when I multiply

01:31

those two I get the area of this little

01:34

rectangle similarly I take f of T 1

01:39

which is that value right here so here's

01:44

F of T 1 and that's this length I

01:48

multiply that beat by t2 minus t1 and

01:51

get the area of the next rectangle and

01:53

so by taking the sum of all these

01:56

products I'm going to add up the values

01:59

of all the areas of these rectangles but

02:03

that's an approximation for the area

02:06

under the curve so in the limit as n

02:09

goes to infinity in these rectangles

02:10

become very narrow they give a very good

02:13

approximation for the area under the

02:15

curve in other words if F happens to be

02:19

above the x-axis just like I'm doing

02:20

right here then this integral is

02:23

actually simply the area under the curve

02:26

otherwise if F is sometimes above and

02:30

sometimes below the x-axis we're going

02:32

to take those portions that are above

02:34

the x-axis and subtract from them the

02:38

portions that are below the x-axis so

02:42

that's the meaning of this integral

02:44

here's the fundamental theorem of

02:47

calculus if f is a function as we

02:50

discussed above we can define a function

02:53

uppercase f of X as the integral from a

02:56

to X F of T DT so in other words if F is

03:00

above the x-axis the integral from a to

03:03

X is simply the area under the curve

03:04

between a and X now X is the variables X

03:08

varies so f of X gives you this area

03:12

depending on that the value of X's of X

03:14

gets bigger in my picture then f of X

03:17

gets bigger if X is smaller f of X gets

03:19

smaller because F happens to be above

03:21

the x-axis but you can define a function

03:24

upper case F like this also if F is

03:26

above and below the x-axis it turns out

03:30

this function that you defined is a

03:32

differentiable function and it's

03:34

derivative is equal to little f of X

03:39

that's the fundamental theorem of

03:41

calculus and why is this so useful upper

03:46

case F actually gives you an

03:49

antiderivative of a function that when

03:52

you differentiate it is little F in

03:56

general we say that a function uppercase

03:58

f that satisfies that a derivative is

04:00

little F such a function is called an

04:03

antiderivative of that another following

04:06

suppose you found two different anti

04:08

derivatives F and G so both of them have

04:11

a property that F prime is equal to

04:15

little F and also G prime is equal to

04:17

the left suppose you found two different

04:19

functions well what is G minus F prime

04:23

what it's G prime minus F prime but both

04:26

of these derivative

04:27

the left so that's zero so G minus F is

04:30

a function whose derivative is zero and

04:33

that makes it a constant function so f

04:35

and G to define anti derivatives of the

04:37

same function f can only differ by a

04:39

constant the conclusion is that we can

04:42

use any anti derivative to evaluate an

04:45

integral and you've probably seen this

04:48

suppose little F is the function you

04:50

want to integrate from A to B and you

04:52

found an antiderivative upper case G and

04:56

the integral from A to B F of T DT is

04:59

the antiderivative evaluated at the

05:01

upper bound B minus the antiderivative

05:04

evaluated at the lower bound a the

05:09

reason for this is if you wanted to

05:10

prove such a fact you would notice that

05:14

G this antiderivative that you found is

05:16

related to the antiderivative F from the

05:19

previous page by simply a constant where

05:23

F was this function f of X is the

05:27

integral from a to X little F of T DT

05:32

which we know what the fundamental

05:34

theorem of calculus that it is an

05:36

antiderivative so what is f of a in that

05:39

case F of a is the integral from a to a

05:42

so that's nothing that's zero and what's

05:44

F of B well that's the full integral of

05:48

actually trying to find so we know this

05:54

is equal to definitely F of B minus F of

05:59

a because F of B is the integral and F

06:02

of a is 0 but now F of B is the same as

06:07

G of B minus this constant C and F of a

06:12

is G of a minus this constant C the

06:16

constants cancel out until you get G of

06:18

B minus G of a so that's how easy that

06:20

is to proof

06:26

so how will we generalize this to see

06:29

instead of having a real valued function

06:32

of a complex valued function and the

06:34

function doesn't have a real variable

06:36

but a complex variable so instead of

06:39

integrating over an interval we're now

06:41

in C so what will we integrate over the

06:45

answer it will integrate over curves

06:47

recall that a curve is a smooth or

06:50

piecewise smooth function gamma to find

06:53

out an interval but into C we could

06:55

write this function as gamma of T equals

06:57

x of T plus I Y of T where X is the real

07:00

part and Y is the imaginary part if F is

07:04

a complex valued function that is

07:07

defined on gamma then we define the

07:10

integral over gamma f of zdz to be the

07:14

limit of these sums we take F of C J

07:18

times the J plus 1 minus CJ and the Z

07:22

J's are the values of gamma at points TJ

07:26

when we divide the integral from A to B

07:29

up into these little chunks again as we

07:31

did before so this is a direct

07:33

generalization of the left hand son

07:35

let's look at the picture and understand

07:37

what the sum really does you can see the

07:41

curve gamma gamma is actually the image

07:46

of the interval from A to B under the

07:52

function gamma again we divide this

07:54

interval up into these little chunks

07:56

where a is equal to t0 B is equal to TN

08:01

and you know this is t1 t2 and so forth

08:08

then we look at the image of these t1 t2

08:12

t3 and so forth so for example the curve

08:14

starts at Z 0 which is gamma of T 0 or

08:18

gamma of a it ends at Zn which is gamma

08:22

of B or gamma of T N and in between we

08:25

have a Z J which is Kamath TJ in the

08:28

next point Z J plus 1 which is gamma of

08:30

TJ plus 1 so we have all these points

08:33

here in order to form this sum we

08:36

evaluate F at ZJ so what's the value of

08:39

F or

08:40

here and multiply that value by Z J plus

08:43

1 minus CJ then we add up all these

08:48

products and take the limit as n goes to

08:51

infinity if that limit exists we call it

08:55

the integral in fact we call it the path

08:58

integral of F over gamma again it's the

09:01

limit as n goes to infinity of the sum F

09:04

of Z J times Z J plus 1 minus Z J when

09:08

the CJS are gamma of T J's and the TJ's

09:11

are those little points that divide the

09:14

integral from A to B up into n pieces

09:17

one can show the following if gamma is a

09:21

smooth curve and F is continuous on

09:24

gamma then this integral f of zdz over

09:28

gamma can be found by taking F evaluated

09:32

at gamma of T multiply that by the

09:35

derivative can my friend of T and take

09:37

the integral from A to B over that

09:40

product why is that here's the idea of

09:44

the proof this is the son that we used

09:47

to define the integral to get the actual

09:49

interval you would put the limit as n

09:52

goes to infinity in front of it but the

09:55

sum of the sum by itself ZJ is simply

10:00

gamma of t j zj + 1 is gamma of t j plus

10:06

1 zj is gamma of t j that makes up this

10:12

sum but now we just divide by ki J plus

10:15

1 minus TJ and then simply multiply by

10:19

TJ plus 1 minus TJ so this will not

10:21

affect the sum at all because we

10:23

multiply and divide by the same number

10:24

but now if you look at this term right

10:28

here as you make TJ plus 1 and TJ be

10:33

closer and closer to each other this

10:35

approaches the derivative gamma prime of

10:40

TJ this is

10:46

approximately the derivative game a

10:47

prime of TJ so the sound is over F of

10:50

gamma TJ times gamma prime of teaching

10:54

times TJ plus 1 minus TJ we often call

10:59

this Delta Chi Chi in the limit this

11:02

goes to the integral from A to B of F of

11:04

gamma of t gamma prime of T DT note we

11:11

haven't really defined how to integrate

11:13

a complex valued function so let's

11:15

quickly look at some examples suppose G

11:19

is a function defined from A to B and

11:21

maps into C so we can write G of T as U

11:24

of T plus IV F T then if we write

11:28

something like integral from A to B G of

11:30

T DT what that really means is the

11:32

interval over u plus I times the

11:35

integral over V in other words we

11:37

integrate the real part we integrate the

11:39

imaginary part and then put them back

11:41

together and that is how the integral

11:43

over G is defined let's look at an

11:45

example suppose I wanted to evaluate the

11:48

integral from 0 to PI e to the I T DT

11:51

you then break up e to the I T as cosine

11:55

T plus I sine T and integrate the cosine

12:00

function and separately integrate the

12:02

sine function an antiderivative for

12:04

cosine T is sine T because the

12:06

derivative of science cousin so we need

12:09

to take sine T and evaluate it at the

12:12

upper bound and subtract from its value

12:14

at the lower bound that's what this

12:16

expresses right here next we need an

12:18

antiderivative for sine T an

12:21

antiderivative for sine T is minus

12:23

cosine T because the derivative of minus

12:25

cosine T is minus minus sine T which is

12:29

sine T so we need to take minus cosine T

12:32

and evaluated at PI and subtract from it

12:34

the value it has at zero while sine of

12:38

PI is actually 0 sine of 0 is also 0 so

12:43

I put that all into this one 0 right

12:45

here cosine of PI is negative 1 cosine

12:51

of 0 is 1 what we're subtracting that

12:53

value so minus another one and here's a

12:55

minus I that's in front of it all

12:57

together this gives you 2

12:59

so the value of this integral is to I

13:08

you actually don't have to break the

13:11

integral up into real and imaginary part

13:14

one can actually alternatively simply

13:16

find an antiderivative right in the

13:18

original form so if you can find an

13:21

antiderivative of e to the I T then you

13:24

can just use the fundamental theorem

13:26

with that antiderivative Y turns out an

13:30

antiderivative of e to the I T is - I

13:33

eat the I T why is that true well let's

13:37

take the derivative - I eat ooh the I T

13:43

I so constant that goes to the side

13:46

and trying to differentiate e to the I T

13:49

the derivative e to the I TS e to the I

13:51

T times the derivative of the inside

13:53

function i T which is I when I multiply

13:58

through I squared is minus 1 together

14:00

with this - it all cancels out so it's e

14:02

to the I T so indeed - I either the I T

14:07

is an antiderivative of either the I T

14:09

so it can take this antiderivative and

14:13

evaluate it at PI and subtract its value

14:16

at 0 the value at pi is minus ie to the

14:20

I pi the value at 0 is minus I e to the

14:24

0 together with the fact that we're

14:26

subtracting it we're getting this plus

14:28

right here e to the I PI is negative 1

14:32

together with this negative sign it's

14:34

plus so this whole expression evaluates

14:39

to I even the 0 is 1 so this also

14:43

evaluates to I so that I end up with 2 I

14:46

so these are two alternative methods to

14:49

find the same integral let's look at

14:52

another example integral from 0 to 1 T

14:55

plus I DT again you could break this up

14:59

into the integral from 0 to 1 of the

15:02

function T DT plus I times the integral

15:06

from 0 to 1 of 1 DT and then find

15:12

antiderivative separately and plug in

15:13

the events or you could just find an

15:16

antiderivative for the whole function

15:19

instead of finding each part separate

15:21

and had derivative 40 is 1/2 T squared

15:24

and a derivative for I is I T and so 1/2

15:29

T squared plus I T is an antiderivative

15:31

for t plus I we need to evaluate that at

15:34

1 and subtract from that the value at 0

15:37

when I plug in T equals 1 I get 1/2 plus

15:40

I when I plug in T equals 0 this all

15:43

vanishes in other words the value of the

15:46

integral is 1/2 plus I now let's look at

15:50

some examples of actual path integrals

15:53

so postive gamma of T is T plus I T for

15:58

T between 0 and 1 what does that look

16:00

like well when T is equal to 0 gamma of

16:06

0 is 0 when T is equal to 1 game up 1 is

16:11

1 plus I 1 and ice right here how about

16:19

in between well I'm graphing X of T and

16:23

y of T together but X of T is equal to Y

16:26

of T so we could write this as X of T

16:29

plus I Y of T and I see that X of T and

16:36

Y of T are always the same value Y is

16:39

equal to X well that's this line right

16:42

here so the path gamma is really the

16:47

line from the origin to 1 plus I we can

16:51

find the derivative as 1 plus I just

16:54

take the derivative of T which is 1 the

16:56

derivative of ITSI and suppose I wanted

16:59

to integrate the function f of Z equals

17:01

Z squared over that path so I need to

17:05

find the integral of f of Z DZ by our

17:09

definition the integral of f of zdz

17:12

because f is a continuous function that

17:14

has smooth I can take f of gamma of t

17:18

times gamma prime of T and evaluate that

17:20

integral from 0 to 1 what is f of gamma

17:24

of T I need to plug in gamma of T into

17:28

my function f but the function f takes

17:31

its input and squares it in other words

17:33

this is gamma of T

17:34

squared game of T squared is T plus I T

17:40

that's gamma of T square so that's where

17:43

this term comes from right here and I

17:45

need to multiply by the derivative of

17:47

gamma but we found that as 1 plus I and

17:52

now I have a integral of a complex

17:56

valued function and I know how to

17:57

evaluate that one way to evaluate that

18:00

is to simply multiply through so this

18:02

first term here becomes T squared plus

18:04

2i times T times t minus T squared the

18:10

minus T squared comes from the I squared

18:12

right here in the second term and I need

18:14

to multiply that by 1 plus I T squared

18:18

and T squared cancel each other out so

18:20

the first term is 2i T squared times 1

18:24

plus I I multiply through I have 2 I

18:27

teeth squared so this term right there

18:30

plus I times 2i T squared but I squared

18:34

is negative 1 so I end up with minus 2t

18:36

squared I could break this up into the

18:40

real part and the imaginary part so I

18:42

get minus 2 times the integral over T

18:45

squared that's from the second term here

18:47

plus 2i times the integral over T

18:50

squared that's the first term and

18:52

antiderivative for T squared is 1/3 T

18:55

cubed so I get minus 2/3 T cubed

18:59

evaluated from 0 to 1 plus 2i and again

19:02

an antiderivative of T squared is 1/3 T

19:05

cubed I'm going to need to evaluate from

19:07

0 to 1 when I plug in 1 for T I find

19:11

minus 2/3 when I plug in 0 this term

19:15

goes away for the second term when I

19:18

plug in T equals 1 I get 2 I over 3 and

19:21

again / 2 equals 0 there is no term

19:23

altogether I can combine these into 2/3

19:26

times negative 1 plus I will be using

19:30

this fact later on again so I'll remind

19:32

you of that when we need

19:39

here's another example suppose I wanted

19:42

to find the integral of 1 over Z DZ over

19:45

the curve for the absolute value of Z

19:48

equals 1 and this is a notation that we

19:51

haven't used so far the absolute value

19:53

of Z being equal to 1 as a set the nose

19:58

the circle of all those Z's were the

20:01

absolute value of Z is once it's the

20:02

circle of radius 1 when we write

20:07

something like that absolute value of z

20:09

equal to 1 we automatically assume that

20:11

the circles oriented this way and most

20:15

of the time the parametrizations gamma

20:18

of T equals either the I T is used for T

20:24

runs from 0 to 2 pi so that's what's

20:29

meant when I write something like this

20:30

integral over absolute value of Z equals

20:33

1 we'll see later that other

20:35

parametrizations feel the same result

20:37

with the integral and that's why this

20:39

notation is actually a reasonable

20:41

notation so how do I find this integral

20:44

if gamma of T is e to the I T then gamma

20:47

prime of T it's the derivative is e to

20:50

the I T times the derivative of the

20:52

inside function which is the exponent so

20:54

times I now the integral over the curve

20:58

gamma of 1 over Z DZ is therefore the

21:00

integral from 0 to 2 pi of 1 over gamma

21:05

of t times gamma prime of T DT so

21:08

whatever we see is e we plug in gamma of

21:11

T and DZ is replaced with gamma prime of

21:14

T DT next let's plug in what gamma of T

21:17

actually is it's either the I T and

21:19

gamma prime of T is ie either the I T

21:21

and we see that the e to the I T term

21:25

cancels out all we're left with this I

21:28

let's pull that outside of the integral

21:30

and were left with the integral from 0

21:32

to 2 pi of DT you could write a 1 right

21:36

here if you wanted to and antiderivative

21:38

of the function 1 is T because the

21:40

derivative of T is 1 we need to evaluate

21:43

that from 0 to 2 pi so we get I times 2

21:47

pi minus I times 0 which is 2 pi I so

21:52

integral of the function 1 over Z over

21:54

the circle of radius 1 is 2 pi here's

21:58

another example same curve is equals 1

22:01

this time we're integrating over the

22:03

function Z instead of 1 over Z we take Z

22:05

so again gamma of T is e to the I T

22:07

gamma prime of T is I in the I T so the

22:12

integral over this curve gamma of Z DZ

22:15

is therefore the integral from 0 to 2pi

22:18

we replace Z gamma and DZ with gamma

22:22

prime of T DT game of t's e to the I T

22:25

gamma prime of keys I eat the I T and we

22:29

can pull this I outside of the integral

22:31

and get the integral from 0 to 2 pi and

22:34

we can combine either the I T and either

22:36

the I T into e to the - I see so now

22:39

what we need is an antiderivative of e

22:42

to the 2i t in the claim is an

22:46

antiderivative of i times e to the 2i t

22:49

is 1/2 times e to the 2i t we better

22:53

check that that is true how do we check

22:56

it well we take the derivative of 1/2 e

22:58

to the 2i t and check said it really is

23:02

what we think it is so what is the

23:04

derivative of 1/2 e to the 2i t 1/2 is a

23:08

constant that goes to this side and the

23:11

derivative of e to the 2i t is e to the

23:14

2i t times the derivative of the

23:18

exponent which is 2i in other words

23:23

these twos cancel out nicely and we're

23:27

left with i times e to the 2i t so

23:30

indeed an antiderivative of i temped e

23:32

to the 2i t is 1/2 e to the 2i t

23:35

and we now take that antiderivative and

23:38

evaluate it at 2 pi subtract from it the

23:40

value at 0 when I plug in 2 pi party I

23:44

get 1/2 times e to the 2 times 2 pi I so

23:48

4 pi I when I plug in 0 I get e to the 0

23:52

but e to the 4 pi i is equal to 1 e to

23:57

the 0 is also equal to 1 and 1 1 is long

24:01

is equal to 0 so the integral evaluates

24:03

to 0

24:10

let's look at the same path gamma and

24:13

integrate 1 over Z squared DZ so so far

24:16

we integrated 1 over Z and we got that

24:23

was 2 pi then we integrated Z to Z and

24:30

we got that was 0 let's see what 1 over

24:32

Z squared gives us well

24:36

again 1 over Z squared means I ride one

24:40

over gamma squared of T or gamma of T

24:42

squared times gamma prime of T DT gamma

24:46

prime of T is this in the numerator and

24:48

it's ie to the I T that's this part

24:50

right here and we divide by the square

24:52

of gamma of G what does gamma of T

24:55

squared that is e to the I T square it's

25:01

either the it1000 I so e to the 2i T's

25:09

here in the denominator and I can

25:12

therefore cancel one of the e to the i

25:14

T's I'm left with 1 e to the I T in the

25:18

denominator or either - I T in the

25:20

numerator what is an antiderivative of I

25:23

II to the - I T the claim is it's minus

25:28

e to the minus I T again let's check how

25:32

do I check I take the derivative of

25:35

minus e to the minus I T the derivative

25:39

of that is minus e to the minus I T

25:42

times the derivative of the exponent

25:44

which is minus I the two negative signs

25:46

cancel out I'm left with items into the

25:49

I T which is exactly what I wanted to

25:51

get so minus e to the minus I t's indeed

25:54

an antiderivative of the function I'm

25:56

integrating and I need to evaluate the

25:58

antiderivative of 2 pi and at 0 at 2 pi

26:01

I get minus e to the minus 2 pi it's 0 I

26:05

get e to the 0 I'm subtracting that but

26:08

with this minus sign I get this plus

26:10

sign the first term is minus 1 either

26:14

the minus 2 pi is 1 with a negative sign

26:16

that gives me a minus 1 e to the 0 is 1

26:19

1 minus 1 0 and again the integral

26:22

it's 2-0 more generally one can see the

26:25

following the integral over Z to the M

26:29

is equal to 2pi when M is equal to minus

26:32

1 when M is equal to minus 1 what I'm

26:39

looking at is the integral over Z equals

26:43

1 of 1 over Z DZ and we showed that that

26:46

was 2pi

26:48

for all other exponents including M

26:51

equals 1 in which case I have the

26:53

integral over Z M equals -2 in which

26:56

case it's the integral 1 over Z squared

26:58

all other cases that integral is equal

27:00

to 0 so all powers of Z the integral

27:04

evaluates to 0 except for the function 1

27:08

over C then we'll see more on why that

27:11

is the case later in the next lecture

27:14

we'll look at more examples and some

27:16

first facts about complex path integrals

Attachments