Chapter 5: DIfferentiation Flashcards

1
Q

DEF: of a limit

A

Let f be defined on some punctured neighbourhood of z0 [i.e., it is defined on {z ∈ C : 0 < |z − z_0| < δ} for some δ > 0]. We use the phrase “f(z) → l as z → z_0” to mean that |f(z)−l| → 0 as |z−z_0| → 0.

Ie f(z) tends to l as z tends to z_0 along any path approaching z_0 and the limit l, does not depend on the path chosen.

Since |f(z)- l | and |z-z_0| are real valued :)

Limit is 2 dimensional: limit must exist in all directions towards z_0

To prove limit does not exist, sufficient to find 2 paths
going to z_0 along which the limits exist and are not equal.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Limit of |z|/z as z tends to 0

A

|z|/z has no limit as z tends to 0

(If exists: limit is 2 dimensional, limit must exist in all directions towards z_0)

|z|/z = 1 for z real and positive
|z|/z = -1 for z real and negative.
Hence there is no l st |z|/z tends to l as z tends to 0

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

INEQUALITIES

A

|u − u₀ |
|v − v₀ |

≤ |w − w₀ | = |u + iv − (u₀ + iv₀ )| ≤ |u − u₀ | + |v − v₀

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

FOR w=u+iv
to tend to
w₀=u₀ +iv₀

as z tends to z₀

(functs of z)

A

for w=u+iv and w₀=u₀ +iv₀

then w tends to w₀ as z tends to z₀ IFF u tends to u₀ AND v tends to v₀ as z tends to ₀

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

DEF 5.2

function of z complex tends to value that f(z₀ ) gives as z tends to z₀

A

If f is defined on a neighbourhood (no longer punctured) of z₀ and
f(z) → f(z₀ ) as z → z₀ , then f is said to be continuous at z₀ .

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

DEF 5.3 f is differentiable at z₀ if …

the derivative of f at z₀ ,

A

Let f be a complex valued function of the complex variable z.
Suppose f is defined on a neighbourhood of z₀.
We say that f is differentiable at z₀ if
lim as z→z₀ of
[(f(z) − f(z₀ )) \ (z − z₀ ) ]
exists.

The limit is the derivative of f at z₀ , denoted f’(z₀).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

DEF 5.4:
analytic

on a region

A

The function f is said to be analytic on a region D if f is differentiable
at all points of D.

analytic for regions (by definition, are open sets)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

DEF 5.5:
analytic

at z₀

A

If z₀ is a point, we say that f is analytic at z₀ if f is analytic on some
region containing z₀ .

  • so if f is analytic at z₀ then its not only differentiable at z₀ but all points within some open disc to non zero radius centred at z₀ .
  • *f is differentiable at z₀ and also at all points sufficiently close to z₀

open sets are natural domains of analytic functions since a function must be defined on a neighbourhood of each point of D

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Familiar results from R which are true in C:

about differentiable function

A

**differentiability implies continuity;

**the sum and product of two differentiable functions is differentiable and the familiar
rules for derivatives of sums and products continue to hold;

**for all non-negative integers n, zⁿ is differentiable and its derivative is nzⁿ⁻¹;

**an analytic function f(w(z)) of an analytic function is analytic on some region and
the chain rule df/dz = (df/dw)(dw/dz) holds.

**if the functions f and g are both differentiable at the point z₀
AND
g(z₀) ≠ 0,

then the quotient f(z)/g(z)
is differentiable at z₀ and its derivative at z₀ is

(f₀’(z₀) g(z₀) − f(z₀) g’(z₀)) /
([g(z₀)]²)

If g(z₀) = 0, then quotient f(z)/g(z) is not defined at z₀ and NOT
differentiable at z₀.

USE THESE TO FIND WHERE:
functions involving quotients are analytic, BY:
firstly finding where they are differentiable.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

EXAMPLE: Let f(z)= Re z

i.e. f(z) = f(x + iy) = x

where is this funct differentiable?

(continuous
everywhere but nowhere differentiable.)

A

Then there is no point of C at
which f is differentiable.

**Firstly let z →z₀ along a line parallel to the real axis

write z₀ = x₀ + iy₀ and z =x+iy₀, where x ≠ x₀.

Then
[f(z) - f(z₀) ] / [ z- z₀] = [x-x₀]\ [x-x₀] = 1.

Thus as z →z₀ along a line parallel to the real axis.
[f(z) - f(z₀) ] / [ z- z₀] → 1

** Secondly, let z →z₀ along a line parallel to the IMAGINARY axis

write z₀ = x₀ + iy₀ and z =x₀+iy, where y ≠ y₀.

Then
[f(z) - f(z₀) ] / [ z- z₀] = [x₀-x₀]\ [i(y-y₀)] = 0.

Thus as z →z₀ along a line parallel to the imaginary axis.
[f(z) - f(z₀) ] / [ z- z₀] → 0

** we see that [f(z) - f(z₀) ] / [ z- z₀] doesn’t tend to and limit as z →z₀.

Hence the function f is not differentiable at z₀.
This is true for all points z₀ ∈C.

** Let f(z) = Re z = u +iv and let z=x+iy. then f(z) = f(x+iy) = x and so u(x,y) =x and v(x,y) =0
 thus 
uₓ≠ v_y at all points of C. Hence Rez isnt analytic on C ( and not differentiable at any point of C) . complex function which
is continuous at every point of C and is not differentiable at any point.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

TMH 5.6 : The Cauchy-Riemann Equations (*)

A

required for functs to be differentiable ie for appropriate limit to exist

Let f(z) = f(x+iy) = u(x, y)+i v(x, y),
where u and v are real valued functions, and let z₀ = x₀ + iy₀. If f is differentiable at z₀,
then u and v satisfy the relations

uₓ = v_y, u_y = −vₓ

at (x₀, y₀),(∗)

for PARTIAL DIFFs

(i.e. ∂u/ ∂x = ∂v/ ∂y ,
∂u/∂y = − ∂v/∂x (∗)
at (x₀, y₀).)

In the complex form; “If the function f is differentiable at z₀, then
i∂f/∂x = ∂f/∂y at z₀.

(*) the two forms are equiv as taking i∂f/∂x =∂f/∂y ,
gives i(uₓ + ivₓ) = u_y + iv_y. Equating real and im gives them
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

PROOF OF

TMH 5.6 : The Cauchy-Riemann Equations (*)

PROOF

A

[ f(z) - f(z₀) ]/ [z-z₀]

As f is differentiable at z₀, f must be defined on a neighbourhood of z₀ and compute df/dz at z₀ by letting z → z₀ in any direction.

**firstly let z → z₀ along a line parallel to the real axis.

write z₀ = x₀ + iy₀ and z =x+iy₀, where x ≠ x₀. Then

( f(z) - f(z₀))/ (z-z₀)
= ([u(x,y₀) + iv(x,y₀)] - [u(x₀,y₀) + iv(x₀,y₀)] ) / x-x₀

= [u(x,y₀) - u(x₀,y₀) ] / x-x₀
+
i [v(x,y₀)- v(x₀,y₀)]/[x-x₀]
→uₓ(x₀,y₀) + ivₓ(x₀,y₀)

as x→x₀

Thus
f’(z₀) = limit as z →z₀

uₓ(x₀,y₀) + ivₓ(x₀,y₀)

**next let z → z₀ along a line parallel to the imaginary axis.

write z₀ = x₀ + iy₀ and z =x₀+iy, where y ≠ y₀.

( f(z) - f(z₀))/ (z-z₀)
= ([u(x₀,y) + iv(x₀,y)] - [u(x₀,y₀) + iv(x₀,y₀)] ) / i(y-y₀)

= [u(x₀,y) - u(x₀,y₀) ] / i(y-y₀)
+
[v(x₀,y)- v(x₀,y₀)]/[ (y-y₀)]

→-iu_y(x₀,y₀) + v_y(x₀,y₀)

as y→y₀

Thus
f’(z₀) = limit as z →z₀

iu_y(x₀,y₀) + v_y(x₀,y₀)

EQUATING REAL AND IMAGINARY PARTS
gives
uₓ(x₀, y₀) = v_y(x₀, y₀), vₓ(x₀, y₀) = −u_y(x₀, y₀)

ie
∂u/∂x = ∂v/∂y ,
∂v/∂x = −∂u/∂y at (x₀,y₀)
ie

f’(z₀)= [uₓ + ivₓ]ₓ₌ₓ₀,y₌y₀ = [∂f/∂x ]ₓ₌ₓ₀,y₌y₀

f'(z₀) =  [-iu_y + v)y]ₓ₌ₓ₀,y₌y₀  
-i = [∂f/∂y ]ₓ₌ₓ₀,y₌y₀ 

HENCE at z₀ = x₀ + iy₀,
∂f/∂x = −i∂f/∂y .

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

PROPERTIES of The Cauchy-Riemann Equations (*)

A
** four versions of f'.
if f= u +iv is analytic
f'(z) 
= uₓ + ivₓ 
= uₓ - iu_y 
=v_y - iu_y
=u_y + ivₓ

** the cauchy- riemann eq are AXIS DOMINATED- only make statements about the appropriate limit in 2 directions, along lines parallel to real and imaginary axes.

(lim as z→z₀ of
[(f(z) − f(z₀ )) \ (z − z₀ ) ] )

We know derivative exists at z₀ if this limit exists for all approaches to it. Hence dont expect the converse to be true.

(there are examples in which the Cauchy-Riemann equations hold at a point
z₀ but the function is not differentiable there.)

the fact that the Cauchy-Riemann equations are satisfied at a certain
point is NOT SUFFICIENT to guarantee differentiability at that point- must be continuous before checking C-R

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q
EXAMPLE: 
z₀ = 0 and let f be defined by 
f(z) =
{1 on the axes
{ 0 elsewhere
 is it continuous/differentiable at the origin
A
write f(z) = u(x,y) + iv(x,y) 
 where  u,v are real valued so that 
v(x,y) = 0, u(x,0)= 1
u(0,y) = 1
for all real x,y. Then

uₓ(0,0) = lim x→x₀ of

[[u(x,0) - u(0,0)]/x] = 0

similarly

u_y(0,0) = 0
vₓ (0,0) = 0
v_y (0,0) = 0
and the C-R equations are satisfied at the origin but f isnt continuous there so not differentiable

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

EXAMPLE:
Prove that, f analytic on C and Re f constant on C implies that f is constant on
C.

A

f(z) = f(x + iy) = u(x, y) + iv(x, y), where u, v are
real-valued. here u = Re f is constant — uₓ = u_y = 0 everywhere.

in aaddition function f is analytic in C and so C-R equ
uₓ = v_y
vₓ = − u_y are satisfied everywhere.

So v is independent of x and y. Thus v is constant, so f = u+iv is constant.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

EXAMPLE:

Prove that, f analytic on C and f’ = 0 on C implies that f is constant on C.

A

f(z) = f(x + iy) = u(x, y) + iv(x, y), where u, v are
real-valued. Here four forms for f’ and since f’ = 0 on C:

0 = f’ = uₓ - i u_y = v_y + ivₓ everywhere.

Hence uₓ = u_y = vₓ = v_y = 0 everywhere and so u and v are constant, so f = u+iv is constant.

17
Q

EXAMPLE:
Let f be analytic on C and suppose that |f| = c, a constant on C. Prove that f is
constant on C.

A

if c= 0 then clearly f(z)=0 for all z∈C

if c bigger than 0 write
f= u +iv. Then

|f(z)|² = [ u(x,y)]² + [v(x,y)]² = c². Taking partial derivatives wrt x and y:

2uuₓ + 2vvₓ = 0
2uu_y + 2vv_y = 0

since f analytic in C, u and v satisfy C-R, so above give:
uuₓ − vu_y = 0
uu_y + vuₓ = 0 .

ELIMINATING u_y
(u² + v²)u²ₓ = 0, i.e. c²u²ₓ = 0 .

hence uₓ = 0 as c is bigger than 0.

Sub back and u_y=0. This u = Re(f) is constant, so f is constant by prev example.

18
Q

non neg powers of z?

A

We have seen that non-negative powers of z can be differentiated at every point of C and
so polynomials are analytic in C.

19
Q

THM 5.7: f is differentiable at z_0 when…

A

Let f(z) = f(x + iy) = u(x, y) + iv(x, y), where u and v are real valued
functions, and let z₀ = x₀ + iy₀. If
(i) u and v have continuous first order partial derivatives at (x₀, y₀)
and
(ii) u and v satisfy the Cauchy-Riemann equations at (x₀, y₀),
then f is differentiable at z₀.

i could also be replaced with condition that u and v are differentiable functs of two vars.
both cond i and ii are necessary and sufficient for differentiability of f

20
Q

SHOW the exponential funct is differentiable everywhere

A
let z = x+iy
eᶻ= eˣ(cosy + isiny) = u(x,y) + iv(x,y)
where
u(x,y) = eˣcosy
v(x,y) = eˣsiny
hence

uₓ = eˣcosy
u_y= −eˣ sin y
vₓ= eˣ sin y
v_y =eˣ cos y

and so first order partial derivatives are continuous and satisfy C-R everywhere:
uₓ = v_y
vₓ= -u_y

exponential funct is therefore differentiable everywhere
and 
d(eᶻ)/dz =   uₓ + ivₓ 
= eˣcosy + eˣ sin y  = eᶻ.

thus the exp funct is ANALYTIC in C.

21
Q
are
cosz 
sinz
coshz
sinhz

analytic everywhere in C

A
Now, for all z,
d/dz (cos z) = d/dz [0.5(eᶦᶻ + e-ᶦᶻ] 
= (-1/2i)(eᶦᶻ - e-ᶦᶻ)
 = − sin z .
Thus cos z is analytic in C and its derivative is − sin z. 
Similarly sin z, cosh z, sinh z are
all analytic in C and they have the familiar derivatives
d/dz (sin z) = cos z,
d/dz (cosh z) = sinh z,
d/dz (sinh z) = cosh z,
everywhere.
22
Q

DEF: harmonic funct

A

A real valued function u(x, y) on a region D ⊆ R² which has

  1. continuous second order partial derivatives, and
  2. which satisfies Laplace’s equation
    ∇²u = uₓₓ + u_yy = 0
    on D

is said to be harmonic on D.

23
Q

THM 5.9: If f = u + iv is analytic on a region D, then u and v

A

If f = u + iv is analytic on a region D, then u and v are harmonic on
D

proof:
Assume uₓ, u_y, vₓ, v_y have second order partial derivatives wrt x and y and that u_xy = u_yx

using C-R eqs:

∇²u =∂²u/∂x²+ ∂²u/∂y²
=∂/∂x (∂u/∂x)  + ∂/∂y (∂u/∂y) 
=∂/∂x (∂v/∂y)  + ∂/∂y (-∂v/∂x) 
= ∂²v/∂x∂y  -  ∂²v/∂y∂x
=0 
∇²v=∂²v/∂x²+ ∂²v/∂y²
=∂/∂x (∂v/∂x)  + ∂/∂y (∂v/∂y) 
=∂/∂x (-∂u/∂y)  + ∂/∂y (∂u/∂x) 
= -∂²u/∂x∂y  +  ∂²u/∂y∂x
=0
24
Q

Given a function u which is harmonic on a region D, can we find a function
f analytic on D with u = Re f?

A

If D is simply connected, then yes.
finding f when u is given:

**METHOD 1 find v by vₓ = - u_y and v_y = u_x. From u +iv and express in terms of z ONLY ( not x and y, Rez, zbar, |z| etc)

**METHOD 2 !!
Form uₓ - iu_y and express in terms of z to get f’(z) and integrate to get f. (v can be ignored)

(one form for f’(z) is this)

25
Q

Example:
u(x, y) = sin(x² + y²) on R²

whether there is a function f analytic on C such that
Re f = u. When f exists, find an expression for f(z) in terms of z.

A

standard: f(z) = f(x + iy) = u(x, y) + iv(x, y), where u, v are
real-valued.

∇²u =∂²/∂x²[ sin(x² + y²)]
\+ ∂²/∂y² [ sin(x² + y²)]
= ∂/∂x [2xcos(x² + y²)] + 
∂/∂y [2ycos(x² + y²)]
= 2cos(x² + y²) - 4x²sin(x² + y²) + 2cos(x² + y²)  - 4y²sin(x² + y²) 
≠ 0
so u is not harmonic and so no f exists
26
Q

Example:

u(x, y) = sinh x cos y −
cosh( x)sin y on R²

whether there is a function f analytic on C such that
Re f = u. When f exists, find an expression for f(z) in terms of z.

A

standard: f(z) = f(x + iy) = u(x, y) + iv(x, y), where u, v are
real-valued.

∇²u =∂²/∂x²[ sinh x cos y −
coshxsin y ]
+ ∂²/∂y² [ sinh x cos y −
cosh( x)sin y ]

= ∂/∂x [ coshxcosy - sinh xsiny]
+ ∂/∂y [ -sinhxsiny - coshxcosy]
=sinhxcosy - coshxsiny - sinhxcosy + coshxsiny
= 0
so u IS harmonic. As C is simply connected we have that f exists,
analytic on C with Re f = u.

using METHOD2 to find f:
f’ =
uₓ - iu_y
= coshxcosy -sinhxsiny + isinhxsiny + icoshxcosy

expressing in terms of z only:
BY CLEVER LITTLE TRICK
f'(x+i0)= (1+i)coshx
so chance that f'(z) = (1+i)coshz
and then 
f(z) = (1+i)sinhz + c for some constant c.

checking: Ref = Re[ (1+i)sinhx + iy) + c]
= Re [(1 + i)(sinh x cosh iy + cosh x sinh iy) + c]
= Re [(1 + i)(sinh x cos y + i cosh x sin y) + c]
= sinh x cos y − cosh x sin y + Re (c)
= u + Re (c)

So f works as long as Re(c) =0

ie
f(z) = (1+i)sinh z +ia
where a ∈ R

all functions f on C with Re f = u are of the form (1 + i) sinh z + ia (a ∈ R)

27
Q

CLEVER LITTLE TRICK

A

sensible guess for f’(z) in terms of z:
writing formula giving f’(x) in terms of x. Then this formula with every x replaced by z.

ie substituting y=0 for f’(x + iy) = f’(x)

28
Q

THM 5.10

properties of f and g if theyre analytic on region D

A

Suppose that f and g are analytic on a region D and Re f = Re g on D.
Then f = g + ia (a ∈ R).

proof:

write h = f - g. Then h is analytic on region D. using standard notation. write
h(z) = h(x+iy) = u(x,y) + iv(x,y) hence,

u(x,y) = Re(h(x+iy)) = Re( f(x+iy)) - Re(g(x+iy)) = 0 on D. Using C-R eq we see v is constant. So h=ia for some a∈ R because Re (h(z)) = 0 on D

29
Q

Final properties for differentiation of f(z)

A

treat f as the main function, and do not split into Re (f) and Im (f).
Analysis in R uses differentiation and integration separately, and unites them in the
Fundamental Theorem of Calculus. There is, however, NOTHING SIMILAR IN COMPLEX
ANALYSIS.

30
Q

Example: The function f is analytic in C and its real and imaginary parts u, v satisfy the
relation u = 1 + v. Show that f is constant.

A

Since u(x,y) = 1+ v(x,y) for all real numbers x,y, differentiating wrt x and wrt y

uₓ = vₓ
u_y=v_y everywhere.
f is analytic in C and so u and v satisfy C-R:
uₓ = v_y,
u_y=-vₓ everywhere.
GIVING:
uₓ = vₓ = −u_y = −v_y = −uₓ everywhere.
hence uₓ = 0 everywhere and thus uₓ =u_y =0 and vₓ = v_y = ₓ everywhere. Thus u and v are constant and so f=u+iv is constant
31
Q

Example: The function f is analytic in C and its real and imaginary parts u and v satisfy
ueᵛ = 12 at all points of C. Prove that f is constant.

A

ueᵛ = 12 everywhere, differentiating wrt x

uₓeᵛ + ueᵛvₓ = 0
-> uₓ + uvₓ =0 as eᵛ≠ 0
differentiating wrt y
u_yeᵛ + ueᵛv_y = 0

rewritten as u_y = uv_y =0.

but f is analytic on C so u and v satisfy C-R eq:
uₓ = v_y,
u_y=-vₓ everywhere.

SUBSTITUTING:
u_y + uuₓ = 0 (9)
uₓ + u(-u_y) =0 (10)

combined give
(1+u²) uₓ = 0

as u is real valued 1 + u² ≥ 1, and so ux = 0 everywhere. Then u_y =0 from (9).

As uₓ = 0 =u_y everywhere u is constant. By C-R vₓ =v_y=0 everywhere so v is also constant. Thus f=u+iv is constant.

32
Q

ftoc

A

none in complex analysis

33
Q

f’ =

A

f’ =

uₓ - iu_y