Joint work with Bob Devaney Small Perturbations of z n Daniel - - PowerPoint PPT Presentation

joint work with bob devaney
SMART_READER_LITE
LIVE PREVIEW

Joint work with Bob Devaney Small Perturbations of z n Daniel - - PowerPoint PPT Presentation

Small Perturbations of z n Daniel Cuzzocreo Boston University Summer Conference on Topology and its Applications Nipissing University July 24, 2013 Small Perturbations of z n Daniel Cuzzocreo ( Boston University) July 24, 2013 1 Joint work


slide-1
SLIDE 1

Small Perturbations of zn

Daniel Cuzzocreo

Boston University

Summer Conference on Topology and its Applications Nipissing University July 24, 2013

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 1

slide-2
SLIDE 2

Joint work with Bob Devaney

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 2

slide-3
SLIDE 3

Basic Notions: Complex Dynamics

In Complex Dynamics, we consider the behavior of points under iteration of a holomorphic function. In this setting, f : ˆ C → ˆ C will be a rational map.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 3

slide-4
SLIDE 4

Basic Notions: Complex Dynamics

In Complex Dynamics, we consider the behavior of points under iteration of a holomorphic function. In this setting, f : ˆ C → ˆ C will be a rational map.

Definitions

For a point z ∈ ˆ C, the sequence (z, f(z), f 2(z), . . . ) is called the

  • rbit of z under f.

If z = f n(z) for some n, with n minimal, then we say z is periodic, with period n. In this case, the complex number λ = (f n)′(z) is called the multiplier of z.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 3

slide-5
SLIDE 5

Julia and Fatou Sets

For a periodic point z, we say z is: attracting if |λ| < 1 superattracting if λ = 0 repelling if |λ| > 1 indifferent if |λ| = 1

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 4

slide-6
SLIDE 6

Julia and Fatou Sets

For a periodic point z, we say z is: attracting if |λ| < 1 superattracting if λ = 0 repelling if |λ| > 1 indifferent if |λ| = 1 This gives a natural partition of the Riemann sphere: The Julia set, J(f), is the closure of the set of repelling periodic points.

Dynamics of f on the Julia set are "chaotic."

The Fatou set is the complement of the Julia set.

Dynamics of f on the Fatou set are "stable."

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 4

slide-7
SLIDE 7

Examples: Quadratic Polynomials

f(z) = z2 "The Unit Circle" f(z) = z2 − 1 "The Basilica" The Julia set is the boundary between the black and orange regions.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 5

slide-8
SLIDE 8

Introduction: Perturbed Polynomials

We consider the singularly perturbed polynomial map Fλ: Fλ(z) = zn + λ zd

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 6

slide-9
SLIDE 9

Introduction: Perturbed Polynomials

We consider the singularly perturbed polynomial map Fλ: Fλ(z) = zn + λ zd Usually n, d ≥ 2. Often (but not always) we take n = d for added symmetry.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 6

slide-10
SLIDE 10

Introduction: Perturbed Polynomials

We consider the singularly perturbed polynomial map Fλ: Fλ(z) = zn + λ zd Usually n, d ≥ 2. Often (but not always) we take n = d for added symmetry. For λ = 0 this map is the complex polynomial z → zn. When λ = 0 we have replaced the superattracting fixed point at the

  • rigin with a pole.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 6

slide-11
SLIDE 11

Motivation

Why study these maps?

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 7

slide-12
SLIDE 12

Motivation

Why study these maps? Allows us to study rational maps of arbitrarily high degree. Many important features in the case n = d = 3, e.g., persist in all higher degrees.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 7

slide-13
SLIDE 13

Motivation

Why study these maps? Allows us to study rational maps of arbitrarily high degree. Many important features in the case n = d = 3, e.g., persist in all higher degrees. As λ → 0, we approach the boundary of Ratn+d, the space of rational maps of degree n + d. The structure of these spaces is a very active area of research.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 7

slide-14
SLIDE 14

Motivation

Why study these maps? Allows us to study rational maps of arbitrarily high degree. Many important features in the case n = d = 3, e.g., persist in all higher degrees. As λ → 0, we approach the boundary of Ratn+d, the space of rational maps of degree n + d. The structure of these spaces is a very active area of research. Symmetries always allow us to study a natural one parameter family in any degree. There is always a single "free" critical orbit.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 7

slide-15
SLIDE 15

Motivation

Why study these maps? Allows us to study rational maps of arbitrarily high degree. Many important features in the case n = d = 3, e.g., persist in all higher degrees. As λ → 0, we approach the boundary of Ratn+d, the space of rational maps of degree n + d. The structure of these spaces is a very active area of research. Symmetries always allow us to study a natural one parameter family in any degree. There is always a single "free" critical orbit. Interesting dynamical behavior and topological features. Sierpi´ nski curve Julia sets are extremely common, for example.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 7

slide-16
SLIDE 16

Motivation

Why study these maps? Allows us to study rational maps of arbitrarily high degree. Many important features in the case n = d = 3, e.g., persist in all higher degrees. As λ → 0, we approach the boundary of Ratn+d, the space of rational maps of degree n + d. The structure of these spaces is a very active area of research. Symmetries always allow us to study a natural one parameter family in any degree. There is always a single "free" critical orbit. Interesting dynamical behavior and topological features. Sierpi´ nski curve Julia sets are extremely common, for example.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 7

slide-17
SLIDE 17

Preliminaries

For this talk, we are interested in the case where |λ| is small. The dynamics here are well understood when n and d are not both 2, but much more complicated when n = d = 2.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 8

slide-18
SLIDE 18

Preliminaries

For this talk, we are interested in the case where |λ| is small. The dynamics here are well understood when n and d are not both 2, but much more complicated when n = d = 2. For simplicity, we’ll assume n = d, so that our map is Fλ(z) = zn + λ zn , n ≥ 2

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 8

slide-19
SLIDE 19

Preliminaries

For this talk, we are interested in the case where |λ| is small. The dynamics here are well understood when n and d are not both 2, but much more complicated when n = d = 2. For simplicity, we’ll assume n = d, so that our map is Fλ(z) = zn + λ zn , n ≥ 2 ∞ is always a superattracting fixed point.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 8

slide-20
SLIDE 20

Preliminaries

For this talk, we are interested in the case where |λ| is small. The dynamics here are well understood when n and d are not both 2, but much more complicated when n = d = 2. For simplicity, we’ll assume n = d, so that our map is Fλ(z) = zn + λ zn , n ≥ 2 ∞ is always a superattracting fixed point. Only pole is at 0, which is also a critical point.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 8

slide-21
SLIDE 21

Preliminaries

For this talk, we are interested in the case where |λ| is small. The dynamics here are well understood when n and d are not both 2, but much more complicated when n = d = 2. For simplicity, we’ll assume n = d, so that our map is Fλ(z) = zn + λ zn , n ≥ 2 ∞ is always a superattracting fixed point. Only pole is at 0, which is also a critical point. 2n other critical points lie at

2n

√ λ.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 8

slide-22
SLIDE 22

Preliminaries

For this talk, we are interested in the case where |λ| is small. The dynamics here are well understood when n and d are not both 2, but much more complicated when n = d = 2. For simplicity, we’ll assume n = d, so that our map is Fλ(z) = zn + λ zn , n ≥ 2 ∞ is always a superattracting fixed point. Only pole is at 0, which is also a critical point. 2n other critical points lie at

2n

√ λ. These map to two critical values at ±2 √ λ

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 8

slide-23
SLIDE 23

Preliminaries

We denote the immediate basin of ∞ by Bλ, and the connected component of the basin of ∞ which contains 0 by Tλ (the "trap door"). These sets may coincide.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 9

slide-24
SLIDE 24

Preliminaries

We denote the immediate basin of ∞ by Bλ, and the connected component of the basin of ∞ which contains 0 by Tλ (the "trap door"). These sets may coincide.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 9

slide-25
SLIDE 25

The Escape Trichotomy

The behavior of the critical points determines the topology of the Julia set of Fλ:

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 10

slide-26
SLIDE 26

The Escape Trichotomy

The behavior of the critical points determines the topology of the Julia set of Fλ:

The Escape Trichotomy (Devaney, Look, Uminsky; 2005)

Let vλ = Fλ(cλ) be a critical value, and suppose F k

λ (cλ) → ∞. Then:

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 10

slide-27
SLIDE 27

The Escape Trichotomy

The behavior of the critical points determines the topology of the Julia set of Fλ:

The Escape Trichotomy (Devaney, Look, Uminsky; 2005)

Let vλ = Fλ(cλ) be a critical value, and suppose F k

λ (cλ) → ∞. Then:

1

if vλ lies in Bλ, then J(Fλ) is a Cantor set;

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 10

slide-28
SLIDE 28

The Escape Trichotomy

The behavior of the critical points determines the topology of the Julia set of Fλ:

The Escape Trichotomy (Devaney, Look, Uminsky; 2005)

Let vλ = Fλ(cλ) be a critical value, and suppose F k

λ (cλ) → ∞. Then:

1

if vλ lies in Bλ, then J(Fλ) is a Cantor set;

2

if vλ lies in Tλ = Bλ, then J(Fλ) is a Cantor set of concentric simple closed curves surrounding the origin;

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 10

slide-29
SLIDE 29

The Escape Trichotomy

The behavior of the critical points determines the topology of the Julia set of Fλ:

The Escape Trichotomy (Devaney, Look, Uminsky; 2005)

Let vλ = Fλ(cλ) be a critical value, and suppose F k

λ (cλ) → ∞. Then:

1

if vλ lies in Bλ, then J(Fλ) is a Cantor set;

2

if vλ lies in Tλ = Bλ, then J(Fλ) is a Cantor set of concentric simple closed curves surrounding the origin;

3

in all other cases, J(Fλ) is a connected set.

In particular, if F j

λ(vλ) ∈ Tλ = Bλ for some j ≥ 1, then J(Fλ) is a

Sierpi´ nski curve (i.e., homeomorphic to the Sierpi´ nski carpet). Sets

  • f λ values where this occurs are called Sierpi´

nski holes.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 10

slide-30
SLIDE 30

Example: Case 1, n=3

A Cantor set.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 11

slide-31
SLIDE 31

Example: Case 2, n=3

A Cantor set of simple closed curves.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 12

slide-32
SLIDE 32

Example: Case 3, n=3

A Sierpi´ nski curve.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 13

slide-33
SLIDE 33

Small perturbations: n ≥ 3

What happens when the parameter is very small? The answer is very different for the cases n = 2 and n ≥ 3.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 14

slide-34
SLIDE 34

Small perturbations: n ≥ 3

What happens when the parameter is very small? The answer is very different for the cases n = 2 and n ≥ 3. When n ≥ 3, there exists a punctured neighborhood of λ = 0 for which Case 2 occurs, (the McMullen domain). The critical values both lie in the trap door and the Julia set is a Cantor set of simple closed curves.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 14

slide-35
SLIDE 35

Small perturbations: n ≥ 3

What happens when the parameter is very small? The answer is very different for the cases n = 2 and n ≥ 3. When n ≥ 3, there exists a punctured neighborhood of λ = 0 for which Case 2 occurs, (the McMullen domain). The critical values both lie in the trap door and the Julia set is a Cantor set of simple closed curves.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 14

slide-36
SLIDE 36

Small perturbations: n ≥ 3

Moreover, as a function of λ, the width of the largest annulus in the Fatou set is bounded away from zero (Devaney, Garijo; 2006).

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 15

slide-37
SLIDE 37

Small perturbations: n ≥ 3

Moreover, as a function of λ, the width of the largest annulus in the Fatou set is bounded away from zero (Devaney, Garijo; 2006). n = 3, λ ≈ −0.005 n = 3, λ ≈ 10−6 There is always at least one "thick" annulus in the Fatou set as λ → 0.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 15

slide-38
SLIDE 38

Small perturbations: n=2

When n = 2, no McMullen domain exists. Why?

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 16

slide-39
SLIDE 39

Small perturbations: n=2

When n = 2, no McMullen domain exists. Why? If the critical values are to lie in the trap door, the Riemann-Hurwitz formula requires that the preimage of the trap door must be a single annulus X mapped 2n-to-1 onto Tλ. The remaining annuli A1 and A2 are each mapped n-to-1

  • nto A = A1 ∪ X ∪ A2.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 16

slide-40
SLIDE 40

Small perturbations: n=2

When n = 2, no McMullen domain exists. Why? An n-to-1 covering map expands the modulus of an annulus by a factor of n, so when n = 2 we have

1 2mod A = mod A1 = mod A2.

But mod A = mod A1 + mod X + mod A2, so there is no room for X.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 17

slide-41
SLIDE 41

Small perturbations, n=2

Since we have no McMullen domain, for small λ, J(Fλ) is connected by the Escape Trichotomy.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 18

slide-42
SLIDE 42

Small perturbations, n=2

Since we have no McMullen domain, for small λ, J(Fλ) is connected by the Escape Trichotomy. Yet there are uncountably many conjugacy classes of maps in any neighborhood of λ = 0.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 18

slide-43
SLIDE 43

Small perturbations, n=2

Since we have no McMullen domain, for small λ, J(Fλ) is connected by the Escape Trichotomy. Yet there are uncountably many conjugacy classes of maps in any neighborhood of λ = 0. Moreover, we have the following theorem:

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 18

slide-44
SLIDE 44

Small perturbations, n=2

Since we have no McMullen domain, for small λ, J(Fλ) is connected by the Escape Trichotomy. Yet there are uncountably many conjugacy classes of maps in any neighborhood of λ = 0. Moreover, we have the following theorem:

Theorem (Devaney, Garijo; 2006)

As λ approaches zero, the Julia set for the map Fλ(z) = z2 + λ z2 converges to the closed unit disk in the Hausdorff metric.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 18

slide-45
SLIDE 45

Small perturbations, n=2

As λ gets small, bounded components of the Fatou set shrink.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 19

slide-46
SLIDE 46

Small perturbations, n=2

Proof Idea

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 20

slide-47
SLIDE 47

Small perturbations, n=2

Proof Idea

If the conclusion fails to hold, then for all ǫ sufficiently small, there exists a sequence λj converging to zero, and a corresponding sequence zj in the closed disk such that Bǫ(zj) lies in the Fatou set of Fλj for all j.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 20

slide-48
SLIDE 48

Small perturbations, n=2

Proof Idea

If the conclusion fails to hold, then for all ǫ sufficiently small, there exists a sequence λj converging to zero, and a corresponding sequence zj in the closed disk such that Bǫ(zj) lies in the Fatou set of Fλj for all j. By compactness, there is a subsequence of the zj converging to a point z∗ in D, so we may assume wlog that Bǫ(z∗) is in the Fatou set of Fλj for all j.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 20

slide-49
SLIDE 49

Small perturbations, n=2

Proof Idea

If the conclusion fails to hold, then for all ǫ sufficiently small, there exists a sequence λj converging to zero, and a corresponding sequence zj in the closed disk such that Bǫ(zj) lies in the Fatou set of Fλj for all j. By compactness, there is a subsequence of the zj converging to a point z∗ in D, so we may assume wlog that Bǫ(z∗) is in the Fatou set of Fλj for all j. Since λj → 0, Fλj ≈ z2 for large j, so for large k, F k

λj wraps Bǫ(z∗)

around the origin, disconnecting the Julia set by forward invariance.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 20

slide-50
SLIDE 50

Parameter space, n=2

Hence the structure of the parameter plane near λ = 0 for n = 2 is quite complicated.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 21

slide-51
SLIDE 51

Parameter space, n=2

Hence the structure of the parameter plane near λ = 0 for n = 2 is quite complicated.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 21

slide-52
SLIDE 52

Parameter space, n=2

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 22

slide-53
SLIDE 53

Parameter space, n=2

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 23

slide-54
SLIDE 54

Parameter space, n=2

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 24

slide-55
SLIDE 55

Parameter space, n=2

Certain structures are clearly visible however:

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 25

slide-56
SLIDE 56

Parameter space, n=2

Certain structures are clearly visible however: We can see many "rings" of Sierpi´ nski holes alternating with what appear to be baby Mandelbrot sets.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 25

slide-57
SLIDE 57

Parameter space, n=2

Certain structures are clearly visible however: We can see many "rings" of Sierpi´ nski holes alternating with what appear to be baby Mandelbrot sets. Looking more closely at any such hole reveals many more such rings surrounding it.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 25

slide-58
SLIDE 58

Parameter space, n=2

Certain structures are clearly visible however: We can see many "rings" of Sierpi´ nski holes alternating with what appear to be baby Mandelbrot sets. Looking more closely at any such hole reveals many more such rings surrounding it. Our goal is ultimately to describe this structure completely.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 25

slide-59
SLIDE 59

Mandelpi´ nski Necklaces

In the dynamical plane, all parameters within the "dividing circle" of radius 1/16 have the property that |vλ| < |cλ|. There exist concentric curves Ck for all integers k such that:

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 26

slide-60
SLIDE 60

Mandelpi´ nski Necklaces

In the dynamical plane, all parameters within the "dividing circle" of radius 1/16 have the property that |vλ| < |cλ|. There exist concentric curves Ck for all integers k such that: C0 is defined to be the critical circle of radius

4

  • |λ|.

Ck+1 surrounds and maps 2-1 onto Ck for k > 0, and C−k lies inside the critical circle and maps 2-1 onto Ck−1 for k > 0.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 26

slide-61
SLIDE 61

Mandelpi´ nski Necklaces

In the dynamical plane, all parameters within the "dividing circle" of radius 1/16 have the property that |vλ| < |cλ|. There exist concentric curves Ck for all integers k such that: C0 is defined to be the critical circle of radius

4

  • |λ|.

Ck+1 surrounds and maps 2-1 onto Ck for k > 0, and C−k lies inside the critical circle and maps 2-1 onto Ck−1 for k > 0. The kth Mandelpi´ nski necklace in the dynamical plane is a simple closed curve of parameters for which the critical values lie on C−k.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 26

slide-62
SLIDE 62

Example: A Mandelpi´ nski Necklace for n=2

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 27

slide-63
SLIDE 63

Sub-necklaces

In the dynamical plane, the sector of points whose arguments lie between two adjacent critical points is mapped 1-1 onto the complement of the rays extending from the critical values to infinity. This sector therefore contains a preimage of the trap door, as well as

  • f all C−k such that vλ lies outside C−k.

Thus the preimage of the trap door is surrounded by infinitely many simple closed curves that map onto these C−k’s. This recursively yields a structure of sub-rings and sub-sub-rings in the dynamical plane that appears to be replicated in parameter space. Can we explicitly show that this structure persists in the parameter plane at each level? (Work in progress).

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 28

slide-64
SLIDE 64

Further Questions

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 29

slide-65
SLIDE 65

Further Questions

Can we explicitly describe the arrangement of all Sierpi´ nski holes in some neighborhood of λ = 0?

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 29

slide-66
SLIDE 66

Further Questions

Can we explicitly describe the arrangement of all Sierpi´ nski holes in some neighborhood of λ = 0? What formula gives the number of Sierpi´ nski holes in each sub-necklace (and sub-sub-necklace, etc.)?

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 29

slide-67
SLIDE 67

Further Questions

Can we explicitly describe the arrangement of all Sierpi´ nski holes in some neighborhood of λ = 0? What formula gives the number of Sierpi´ nski holes in each sub-necklace (and sub-sub-necklace, etc.)? Can we prove the existence of all baby Mandelbrot sets in each necklace?

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 29

slide-68
SLIDE 68

Thanks

Thank you for your attention.

Daniel Cuzzocreo ( Boston University) Small Perturbations of zn July 24, 2013 30