Connectedness properties of the set where the iterates of an entire - - PowerPoint PPT Presentation
Connectedness properties of the set where the iterates of an entire - - PowerPoint PPT Presentation
Connectedness properties of the set where the iterates of an entire function are unbounded John Osborne (joint work with Phil Rippon and Gwyneth Stallard) Postgraduate Conference in Complex Dynamics 11 - 13 March 2015 The set of points whose
The set of points whose orbits are unbounded
f an entire function, f n = f ◦ f ◦ . . . ◦ f (f n(z))n∈N the orbit of z under f K(f) = set of points whose orbits are bounded So K(f)c = set of points whose orbits are unbounded
The set of points whose orbits are unbounded
f an entire function, f n = f ◦ f ◦ . . . ◦ f (f n(z))n∈N the orbit of z under f K(f) = set of points whose orbits are bounded So K(f)c = set of points whose orbits are unbounded K(f)c ⊃ I(f)
The set of points whose orbits are unbounded
f an entire function, f n = f ◦ f ◦ . . . ◦ f (f n(z))n∈N the orbit of z under f K(f) = set of points whose orbits are bounded So K(f)c = set of points whose orbits are unbounded K(f)c ⊃ I(f) f a polynomial K(f) = filled Julia set K(f)c = I(f) ⊂ F(f) f transcendental K(f) unbounded K(f)c \ I(f) = ∅
The set of points whose orbits are unbounded
f an entire function, f n = f ◦ f ◦ . . . ◦ f (f n(z))n∈N the orbit of z under f K(f) = set of points whose orbits are bounded So K(f)c = set of points whose orbits are unbounded K(f)c ⊃ I(f) f a polynomial K(f) = filled Julia set K(f)c = I(f) ⊂ F(f) K(f)c is connected f transcendental K(f) unbounded K(f)c \ I(f) = ∅ When is K(f)c connected?
Iterating the minimum modulus function
Define m(r) = m(r, f) := min{|f(z)| : |z| = r} mn(r) to be the nth iterate of the function r → m(r). r f
Iterating the minimum modulus function
Define m(r) = m(r, f) := min{|f(z)| : |z| = r} mn(r) to be the nth iterate of the function r → m(r). r f m(r)
Iterating the minimum modulus function
Define m(r) = m(r, f) := min{|f(z)| : |z| = r} mn(r) to be the nth iterate of the function r → m(r). r f m(r) f m2(r)
Iterating the minimum modulus function
Define m(r) = m(r, f) := min{|f(z)| : |z| = r} mn(r) to be the nth iterate of the function r → m(r). r f m(r) f m2(r) f m3(r)
Iterating the minimum modulus function
Define m(r) = m(r, f) := min{|f(z)| : |z| = r} mn(r) to be the nth iterate of the function r → m(r). r f m(r) f m2(r) f m3(r) For a transcendental entire function (compare the iteration
- f M(r)):
∄ R > 0 with m(r) > r ∀ r ≥ R
Iterating the minimum modulus function
Define m(r) = m(r, f) := min{|f(z)| : |z| = r} mn(r) to be the nth iterate of the function r → m(r). r f m(r) f m2(r) f m3(r) For a transcendental entire function (compare the iteration
- f M(r)):
∄ R > 0 with m(r) > r ∀ r ≥ R we can’t always find r > 0 such that mn(r) → ∞ as n → ∞.
Some functions for which K(f)c is connected
Theorem A Let f be a transcendental entire function for which there exists r > 0 such that mn(r) → ∞ as n → ∞. Then K(f)c is connected.
Some functions for which K(f)c is connected
Theorem A Let f be a transcendental entire function for which there exists r > 0 such that mn(r) → ∞ as n → ∞. Then K(f)c is connected. Theorem B Let f be a transcendental entire function of order less than 1
2.
Then there exists r > 0 such that mn(r) → ∞ as n → ∞, and therefore K(f)c is connected. Recall that the order ρ of a transcendental entire function is defined as ρ := lim sup
r→∞
log log M(r, f) log r .
An idea of the proof of Theorem A
Suppose K(f)c is disconnected. Lemma A subset X of C is disconnected if and only if there exists a closed, connected set Γ ⊂ X c such that at least two different components of Γc intersect X.
An idea of the proof of Theorem A
Suppose K(f)c is disconnected. Lemma A subset X of C is disconnected if and only if there exists a closed, connected set Γ ⊂ X c such that at least two different components of Γc intersect X. mk(r) z0 z1 Γ ⊂ K(f)
An idea of the proof of Theorem A
Suppose we have a continuum Γ0 ⊂ K(f) such that for some z0 ∈ Γ0, |f n(z0)| < mk(r) for all n ∈ N, and ∃ z1 ∈ Γ0 ∩ {z : |z| = mk(r)}. mk(r) Γ0 z0 z1
An idea of the proof of Theorem A
Suppose we have a continuum Γ0 ⊂ K(f) such that for some z0 ∈ Γ0, |f n(z0)| < mk(r) for all n ∈ N, and ∃ z1 ∈ Γ0 ∩ {z : |z| = mk(r)}. mk(r) Γ0 z0 z1 mk+N(r) f N(Γ0) f N(z0) f N(z1) mk+N+1(r) N ≥ 1 is the largest integer such that |f N(z1)| ≥ mk+N(r).
An idea of the proof of Theorem A
Suppose we have a continuum Γ0 ⊂ K(f) such that for some z0 ∈ Γ0, |f n(z0)| < mk(r) for all n ∈ N, and ∃ z1 ∈ Γ0 ∩ {z : |z| = mk(r)}. mk(r) Γ0 z0 z1 mk+N(r) f N(z1) mk+N+1(r) z2 Γ1 Choose Γ1 ⊂ f N(Γ0) so that it contains a point z2 with modulus mk+N(r) but no points
- f smaller modulus.
An idea of the proof of Theorem A
Suppose we have a continuum Γ0 ⊂ K(f) such that for some z0 ∈ Γ0, |f n(z0)| < mk(r) for all n ∈ N, and ∃ z1 ∈ Γ0 ∩ {z : |z| = mk(r)}. mk(r) Γ0 z0 z1 mk+N(r) f N(z1) mk+N+1(r) z2 Γ1 mk+N+L(r) f L(Γ1) f N+L(z1) f L(z2) L ≥ 1 is the largest integer such that |f L(z2)| ≥ mk+N+L(r).
An idea of the proof of Theorem A
Suppose we have a continuum Γ0 ⊂ K(f) such that for some z0 ∈ Γ0, |f n(z0)| < mk(r) for all n ∈ N, and ∃ z1 ∈ Γ0 ∩ {z : |z| = mk(r)}. mk(r) Γ0 z0 z1 mk+N(r) mk+N+1(r) z2 Γ1 mk+N+L(r) z3 Γ2 Choose Γ2 ⊂ f L(Γ1) so that it contains a point z3 with modulus mk+N+L(r) but no points of smaller modulus.
An idea of the proof of Theorem A
We have constructed a sequence (Γn) of compact sets such that f kn(Γn) ⊃ Γn+1 for some (kn). mk(r) Γ0 mk+N(r) mk+N+1(r) Γ1 mk+N+L(r) Γ2 f N f L
An idea of the proof of Theorem A
We have constructed a sequence (Γn) of compact sets such that f kn(Γn) ⊃ Γn+1 for some (kn). mk(r) Γ0 mk+N(r) mk+N+1(r) Γ1 mk+N+L(r) Γ2 f N f L It follows that there is a point in Γ0 with unbounded orbit. #
Generalising the condition in Theorem A
‘... there exists r > 0 such that mn(r) → ∞ as n → ∞.’ r f m(r) We have: a sequence of nested discs {z : |z| < mn(r)} that fill the plane such that each boundary circle is mapped outside the next disc in the sequence. Can we replace the discs by arbitrary bounded, simply connected domains?
A more general result
Theorem C Let f be a transcendental entire function, and (Dn)n∈N be a sequence of bounded, simply connected domains such that (a) f(∂Dn) surrounds Dn+1, for n ∈ N, and (b) every disc centred at 0 is contained in Dn for sufficiently large n. Then K(f)c is connected.
A more general result
Theorem C Let f be a transcendental entire function, and (Dn)n∈N be a sequence of bounded, simply connected domains such that (a) f(∂Dn) surrounds Dn+1, for n ∈ N, and (b) every disc centred at 0 is contained in Dn for sufficiently large n. Then K(f)c is connected. Is this really more general than Theorem A?
A more general result
Theorem C Let f be a transcendental entire function, and (Dn)n∈N be a sequence of bounded, simply connected domains such that (a) f(∂Dn) surrounds Dn+1, for n ∈ N, and (b) every disc centred at 0 is contained in Dn for sufficiently large n. Then K(f)c is connected. Is this really more general than Theorem A? Example: Let f(z) = −10ze−z − 1
2z.
Note that m(r) ∼ 1
2r as r → ∞, so Theorem A does not hold.
Example: f(z) = −10ze−z − 1
2z
4nπ nπi −nπi
∂Dn
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f a
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f b a
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f f b a
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f c f b a
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f e c f b a
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f d d e c f b a
Example: f(z) = −10ze−z − 1
2z
4(n + 1)π 4nπ nπi −nπi 4(n + 1)πi −4(n + 1)πi
∂Dn ∂Dn+1
a b c d e f d d e c f b a