Bounded Hyperbolic Components of Bicritical Rational Maps Hongming - - PowerPoint PPT Presentation

bounded hyperbolic components of bicritical rational maps
SMART_READER_LITE
LIVE PREVIEW

Bounded Hyperbolic Components of Bicritical Rational Maps Hongming - - PowerPoint PPT Presentation

Bounded Hyperbolic Components of Bicritical Rational Maps Hongming Nie (joint with K. Pilgrim) The Hebrew University of Jerusalem Topics in Complex Dynamics 2019 Barcelona, Spain Definitions A complex rational map f : P 1 P 1 is


slide-1
SLIDE 1

Bounded Hyperbolic Components of Bicritical Rational Maps

Hongming Nie (joint with K. Pilgrim)

The Hebrew University of Jerusalem

Topics in Complex Dynamics 2019 Barcelona, Spain

slide-2
SLIDE 2

Definitions

◮ A complex rational map f : P1 → P1 is hyperbolic if each

critical point converges under iteration to an attracting cycle.

slide-3
SLIDE 3

Definitions

◮ A complex rational map f : P1 → P1 is hyperbolic if each

critical point converges under iteration to an attracting cycle.

◮ The set of hyperbolic rational maps of degree d ≥ 2 is an open

set in the space Ratd of degree d rationals maps. It descends an open set in the muduli space ratd := Ratd/Aut(P1).

slide-4
SLIDE 4

Definitions

◮ A complex rational map f : P1 → P1 is hyperbolic if each

critical point converges under iteration to an attracting cycle.

◮ The set of hyperbolic rational maps of degree d ≥ 2 is an open

set in the space Ratd of degree d rationals maps. It descends an open set in the muduli space ratd := Ratd/Aut(P1).

◮ Each component of the set of hyperbolic maps is called a

hyperbolic component.

slide-5
SLIDE 5

Definitions

◮ A complex rational map f : P1 → P1 is hyperbolic if each

critical point converges under iteration to an attracting cycle.

◮ The set of hyperbolic rational maps of degree d ≥ 2 is an open

set in the space Ratd of degree d rationals maps. It descends an open set in the muduli space ratd := Ratd/Aut(P1).

◮ Each component of the set of hyperbolic maps is called a

hyperbolic component.

◮ type D hyperbolic component: each map has maximal number

  • f disjoint attracting cycles.

strict type D hyperbolic component: type D + each attracting cycle has period at least 2.

slide-6
SLIDE 6

Bounded hyperbolic components

Let V ⊂ ratd be a subvariety. We say a hyperbolic component H ⊂ V is bounded if its closure H is compact in V .

slide-7
SLIDE 7

Bounded hyperbolic components

Let V ⊂ ratd be a subvariety. We say a hyperbolic component H ⊂ V is bounded if its closure H is compact in V .

Theorem (Epstein, ’00)

Let H be a strict type D hyperbolic component in rat2. Then H is bounded.

slide-8
SLIDE 8

Epstein’s argument

Suppose H is unbounded.

slide-9
SLIDE 9

Epstein’s argument

Suppose H is unbounded.

◮ Milnor ’93: rat2 ∼

= C2, and a sequence [fk] is unbounded in rat2 if and only if at least one multiplier of a fixed point tends to ∞.

slide-10
SLIDE 10

Epstein’s argument

Suppose H is unbounded.

◮ Milnor ’93: rat2 ∼

= C2, and a sequence [fk] is unbounded in rat2 if and only if at least one multiplier of a fixed point tends to ∞.

◮ Do analytic estimates on the three multipliers to obtain limit

dynamics.

slide-11
SLIDE 11

Epstein’s argument

Suppose H is unbounded.

◮ Milnor ’93: rat2 ∼

= C2, and a sequence [fk] is unbounded in rat2 if and only if at least one multiplier of a fixed point tends to ∞.

◮ Do analytic estimates on the three multipliers to obtain limit

dynamics.

◮ Analyze the limits of the two attracting cycles.

slide-12
SLIDE 12

Epstein’s argument

Suppose H is unbounded.

◮ Milnor ’93: rat2 ∼

= C2, and a sequence [fk] is unbounded in rat2 if and only if at least one multiplier of a fixed point tends to ∞.

◮ Do analytic estimates on the three multipliers to obtain limit

dynamics.

◮ Analyze the limits of the two attracting cycles. ◮ Get a contradiction with the limit dynamics.

slide-13
SLIDE 13

Epstein’s argument

Suppose H is unbounded.

◮ Milnor ’93: rat2 ∼

= C2, and a sequence [fk] is unbounded in rat2 if and only if at least one multiplier of a fixed point tends to ∞.

◮ Do analytic estimates on the three multipliers to obtain limit

dynamics.

◮ Analyze the limits of the two attracting cycles. ◮ Get a contradiction with the limit dynamics.

It seems not easy to reproduce this argument for rational maps of higher degree.

slide-14
SLIDE 14

Bicritical Rational Maps

◮ A rational map is bicritical if it has exact two critical points.

slide-15
SLIDE 15

Bicritical Rational Maps

◮ A rational map is bicritical if it has exact two critical points. ◮ By conjugating so that the two critical points are at 0 and ∞

and a fixed point is at 1. F := αzd + β γzd + δ : αδ − βγ = 1, α + β = γ + δ

  • ⊂ Ratd.
slide-16
SLIDE 16

Bicritical Rational Maps

◮ A rational map is bicritical if it has exact two critical points. ◮ By conjugating so that the two critical points are at 0 and ∞

and a fixed point is at 1. F := αzd + β γzd + δ : αδ − βγ = 1, α + β = γ + δ

  • ⊂ Ratd.

◮ Choose suitable coordinates so that

F = C2 − {2 lines} ⊂ C2 ⊂ P2.

slide-17
SLIDE 17

Bicritical Rational Maps

◮ A rational map is bicritical if it has exact two critical points. ◮ By conjugating so that the two critical points are at 0 and ∞

and a fixed point is at 1. F := αzd + β γzd + δ : αδ − βγ = 1, α + β = γ + δ

  • ⊂ Ratd.

◮ Choose suitable coordinates so that

F = C2 − {2 lines} ⊂ C2 ⊂ P2.

◮ Let Md be the moduli space of bicritical rational maps of

degree d. Then a hyperbolic component H ⊂ Md lifts to a hyperbolic component H ⊂ F.

slide-18
SLIDE 18

Main Result

Theorem (N.-Pilgrim)

Let H ⊂ Md be a strict type D hyperbolic component. Then H is bounded in Md.

slide-19
SLIDE 19

Sketch of proof

Accessibility of ideal points:

slide-20
SLIDE 20

Sketch of proof

Accessibility of ideal points:

◮ The lift

H of H is semi-algebraic (Milnor ’14).

slide-21
SLIDE 21

Sketch of proof

Accessibility of ideal points:

◮ The lift

H of H is semi-algebraic (Milnor ’14).

◮ Curve Section Lemma ⇒ any boundary point of

H in P2 can be approached by a sequence from a holomorphic family.

slide-22
SLIDE 22

Sketch of proof

Accessibility of ideal points:

◮ The lift

H of H is semi-algebraic (Milnor ’14).

◮ Curve Section Lemma ⇒ any boundary point of

H in P2 can be approached by a sequence from a holomorphic family. In summary, if H is unbounded, we can find a holomorphic family {ft}t∈D∗ ⊂ F such that for some tk → 0, ftk ∈ H and [ftk] → ∞ in ratd.

slide-23
SLIDE 23

Sketch of proof

Accessibility of ideal points:

◮ The lift

H of H is semi-algebraic (Milnor ’14).

◮ Curve Section Lemma ⇒ any boundary point of

H in P2 can be approached by a sequence from a holomorphic family. In summary, if H is unbounded, we can find a holomorphic family {ft}t∈D∗ ⊂ F such that for some tk → 0, ftk ∈ H and [ftk] → ∞ in ratd. From now on, we assume H is unbound and consider the family {ft}.

slide-24
SLIDE 24

Sketch of proof (cont.)

Induced map on Berkovich space (following Kiwi ’15):

slide-25
SLIDE 25

Sketch of proof (cont.)

Induced map on Berkovich space (following Kiwi ’15):

◮ The holomorphic family {ft} induces a rational map

f(z) ∈ C((t))(z) ⊂ C{{t}}(z) ⊂ L(z), where C((t)) is the field of Laurent series, C{{t}} is the field

  • f Puiseux series, and L is the completion of C{{t}} w.r.t the

natural non-Archimedean absolute value.

slide-26
SLIDE 26

Sketch of proof (cont.)

Induced map on Berkovich space (following Kiwi ’15):

◮ The holomorphic family {ft} induces a rational map

f(z) ∈ C((t))(z) ⊂ C{{t}}(z) ⊂ L(z), where C((t)) is the field of Laurent series, C{{t}} is the field

  • f Puiseux series, and L is the completion of C{{t}} w.r.t the

natural non-Archimedean absolute value.

◮ The map f extends to an endomorphism on Berkovich space

P1 over L. (The Berkovich space P1 is a compact, Hausdorff, uniquely path-connected topological space with tree structure.)

slide-27
SLIDE 27

Sketch of proof (cont.)

Figure 1: The Berkovich space P1 (see book “Berkovich Spaces and Applications”)

slide-28
SLIDE 28

Sketch of proof (cont.)

Berkovich dynamics of f (bicritical rational map. The quadratic case was done by Kiwi ’14):

slide-29
SLIDE 29

Sketch of proof (cont.)

Berkovich dynamics of f (bicritical rational map. The quadratic case was done by Kiwi ’14):

◮ The two cycles zt and wt of ft induce two non-repelling

cycles z and w of f.

slide-30
SLIDE 30

Sketch of proof (cont.)

Berkovich dynamics of f (bicritical rational map. The quadratic case was done by Kiwi ’14):

◮ The two cycles zt and wt of ft induce two non-repelling

cycles z and w of f.

◮ It follows that f has a repelling q-cycle for some q ≥ 2 where

the reduction G of fq is a degree d bicritical rational map with a multiple fixed point ˆ z.

slide-31
SLIDE 31

Sketch of proof (cont.)

Berkovich dynamics of f (bicritical rational map. The quadratic case was done by Kiwi ’14):

◮ The two cycles zt and wt of ft induce two non-repelling

cycles z and w of f.

◮ It follows that f has a repelling q-cycle for some q ≥ 2 where

the reduction G of fq is a degree d bicritical rational map with a multiple fixed point ˆ z.

◮ The limit of the cycle zt ( resp. of wt) is either {ˆ

z}, contains a cycle disjoint from ˆ z, or contains a preperiodic critical point that iterates under G to ˆ z.

slide-32
SLIDE 32

Sketch of proof (cont.)

Contradiction:

slide-33
SLIDE 33

Sketch of proof (cont.)

Contradiction: Applying

◮ an arithmetic result of Rivera-Letelier: number of fixed points

in a Berkovich Fatou component.

◮ Epstein’s refined version of the Fatou-Shishikura Inequality:

relations on the numbers of critical points and non-repelling cycles.

slide-34
SLIDE 34

Sketch of proof (cont.)

Contradiction: Applying

◮ an arithmetic result of Rivera-Letelier: number of fixed points

in a Berkovich Fatou component.

◮ Epstein’s refined version of the Fatou-Shishikura Inequality:

relations on the numbers of critical points and non-repelling cycles. We derive an over-determined set of constraints on the critical dynamics of G.

slide-35
SLIDE 35

Thank you.