23.11.2012 Views

Discrete Holomorphic Local Dynamical Systems

Discrete Holomorphic Local Dynamical Systems

Discrete Holomorphic Local Dynamical Systems

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

218 Tien-Cuong Dinh and Nessim Sibony<br />

Let Λg denote the adjoint of the operator ϕ ↦→ ϕ ◦ g on L 2 (ν). These operators<br />

are of norm 1. The computation in Lemma 1.96 shows that E(ϕ|F1)=Λg(ϕ) ◦ g.<br />

We obtain in the same way that E(ϕ|Fn)=Λ n g (ϕ) ◦ g n .Define<br />

ψ ′′ := −<br />

∞<br />

∑<br />

n=1<br />

Λ n g (ψ), ψ′ := ψ − (ψ ′′ − ψ ′′ ◦ g).<br />

Using the hypotheses in Theorem 1.94, we see that ψ ′ and ψ ′′ are in L 2 (ν) with<br />

norms bounded by some constant. However, we loose the uniform boundedness:<br />

these functions are not necessarily in L ∞ (ν).<br />

Lemma 1.97. We have Λ n g (ψ ′ )=0 for n ≥ 1 and E(ψ ′ ◦ g n |Fm)=0 for m > n ≥ 0.<br />

Proof. Clearly Λg(ψ ′′ ◦ g)=ψ ′′ . We deduce from the definition of ψ ′′ that<br />

Λg(ψ ′ )=Λg(ψ) − Λg(ψ ′′ )+Λg(ψ ′′ ◦ g)=Λg(ψ) − Λg(ψ ′′ )+ψ ′′ = 0.<br />

Hence, Λ n g (ψ′ )=0forn ≥ 1. For every function φ in L 2 (ν), sinceν is invariant,<br />

we have for m > n<br />

〈ν,(ψ ′ ◦ g n )(φ ◦ g m )〉 = 〈ν,ψ ′ (φ ◦ g m−n )〉 = 〈ν,Λ m−n<br />

g (ψ ′ )φ〉 = 0.<br />

It follows that E(ψ ′ ◦ g n |Fm)=0. ⊓⊔<br />

Lemma 1.98. There are constants δ0 > 1 and c > 0 such that<br />

ν{|ψ ′ | > b}≤ce −δ b 0 and ν{|ψ ′′ | > b}≤ce −δ b 0<br />

for any b ≥ 0. In particular, tψ ′ and tψ ′′ are ν-integrable for every t ≥ 0.<br />

Proof. Since ψ ′ := ψ − (ψ ′′ − ψ ′′ ◦ g) and ψ is bounded, it is enough to prove the<br />

estimate on ψ ′′ . Indeed, the invariance of ν implies that ψ ′′ ◦ g satisfies a similar<br />

inequality.<br />

Fix a positive constant δ1 such that 1 < δ 2 1<br />

< δ, whereδ is the constant<br />

in Theorem 1.94. Define ϕ := ∑n≥1 δ 2n<br />

1 |Λ n g (ψ)|. We first show that there is<br />

a constant α > 0 such that ν{ϕ ≥ b} � e −αb for every b ≥ 0. Recall that<br />

E(ψ|Fn) =Λ n g (ψ) ◦ gn . Using the hypothesis of Theorem 1.94, the inequality<br />

∑ 1<br />

2n 2 ≤ 1 and the invariance of ν, we obtain for b ≥ 0<br />

ν{ϕ ≥ b} ≤∑ ν<br />

n≥1<br />

= ∑ ν<br />

n≥1<br />

�<br />

�<br />

|Λ n g (ψ)|≥<br />

−2n<br />

δ1 2n2 �<br />

b<br />

δ n |E(ψ|Fn)|≥ δ nδ −2n<br />

1<br />

2n2 ≤ ∑ ν<br />

n≥1<br />

�<br />

b<br />

�<br />

|E(ψ|Fn)|≥<br />

�<br />

� ∑ exp<br />

n≥1<br />

It follows that ν{ϕ ≥ b} � e −αb for some constant α > 0.<br />

−2n<br />

δ1 2n2 −δ nδ −2n<br />

1<br />

2n2 �<br />

b<br />

�<br />

b<br />

.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!