Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
MATH41112/61112 Ergodic Theory Lecture 21 21. Birkhoff ’s Ergodic Theorem §21.1 Introduction An ergodic theorem is a result that describes the limiting behaviour of the sequence n−1 1X f ◦ Tj (21.1) n j=0 as n → ∞. The precise formulation of an ergodic theorem depends on the class of function f (for example, one could assume that f is integrable, L2 , or continuous), and the notion of convergence that we use (for example, we could study pointwise convergence, L 2 convergence, or uniform convergence). The result that we are interested here—Birkhoff’s Ergodic Theorem—deals with pointwise convergence of (21.1) for an integrable function f . §21.2 Conditional expectation We will need the concepts of Radon-Nikodym derivates and conditional expectation. Definition. Let µ be a measure on (X, B). We say that a measure ν is absolutely continuous with respect to µ and write ν µ if ν(B) = 0 whenever µ(B) = 0, B ∈ B. Remark. Thus ν is absolutely continuous with respect to µ if sets of µmeasure zero also have ν-measure zero (but there may be more sets of νmeasure zero). For example, let f ∈ L1 (X, B, µ) be non-negative and define a measure ν by Z ν(B) = f dµ. B Then ν µ. The following theorem says that, essentially, all absolutely continuous measures occur in this way. Theorem 21.1 (Radon-Nikodym) Let (X, B, µ) be a probability space. Let ν be a measure defined on B and suppose that ν µ. Then there is a non-negative measurable function f such that Z f dµ, for all B ∈ B. ν(B) = B 1 MATH41112/61112 Ergodic Theory Lecture 21 Moreover, f is unique in the sense that if g is a measurable function with the same property then f = g µ-a.e. Exercise 21.1 If ν µ then it is customary to write dν/dµ for the function given by the Radon-Nikodym theorem, that is Z dν ν(B) = dµ. B dµ Prove the following relations: (i) If ν µ and f is a µ-integrable function then Z Z dν dµ. f dν = f dµ (ii) If ν1 , ν2 µ then d(ν1 + ν2 ) dν1 dν2 = + . dµ dµ dµ (iii) If λ ν µ then dλ dν dλ = . dµ dν dµ Let A ⊂ B be a sub-σ-algebra. Note that µ defines a measure on A by restriction. Let f ∈ L1 (X, B, µ). Then we can define a measure ν on A by setting Z f dµ. ν(A) = A Note that ν µ|A . Hence by the Radon-Nikodym theorem, there is a unique A-measurable function E(f | A) such that Z ν(A) = E(f | A) dµ. We call E(f | A) the conditional expectation of f with respect to the σalgebra A. So far, we have only defined E(f | A) for non-negative f . To define E(f | A) for an arbitrary f , we split f into positive and negative parts f = f+ − f− where f+ , f− ≥ 0 and define E(f | A) = E(f+ | A) − E(f− | A). Thus we can view conditional expectation as an operator E(· | A) : L1 (X, B, µ) → L1 (X, A, µ). Note that E(f | A) is uniquely determined by the two requirements that 2 MATH41112/61112 Ergodic Theory Lecture 21 (i) E(f | A) is A-measurable, and R R (ii) A f dµ = A E(f | A) dµ for all A ∈ A. Intuitively, one can think of E(f | A) as the best approximation to f in the smaller space of all A-measurable functions. Exercise 21.2 (i) Prove that f 7→ E(f | A) is linear. (ii) Suppose that g is A-measurable and |g| < ∞ µ-a.e. Show that E(f g | A) = gE(f | A). (iii) Suppose that T is a measure-preserving transformation. Show that E(f | A) ◦ T = E(f ◦ T | T −1 A). (iv) Show that E(f | B) = f . (v) Let N denote the trivial σ-algebra consisting of all sets of measure 0 R and 1. Show that E(f | N ) = f dµ. To state Birkhoff’s Ergodic Theorem precisely, we will need the sub-σalgebra I of T -invariant subsets, namely: I = {B ∈ B | T −1 B = B a.e.}. Exercise 21.3 Prove that I is a σ-algebra. §21.3 Birkhoff ’s Pointwise Ergodic Theorem Birkhoff’s Ergodic Theorem deals with the behaviour of µ-a.e. x ∈ X, and for f ∈ L1 (X, B, µ). 1 n Pn−1 j=0 f (T j x) for Theorem 21.2 (Birkhoff ’s Ergodic Theorem) Let (X, B, µ) be a probability space and let T : X → X be a measurepreserving transformation. Let I denote the σ-algebra of T -invariant sets. Then for every f ∈ L1 (X, B, µ), we have n−1 1X f (T j x) → E(f | I) n j=0 for µ-a.e. x ∈ X. 3 MATH41112/61112 Ergodic Theory Lecture 21 Corollary 21.3 Let (X, B, µ) be a probability space and let T : X → X be an ergodic measure-preserving transformation. Let f ∈ L 1 (X, B, µ). Then n−1 1X f (T j x) → n j=0 Z f dµ, as n → ∞, for µ-a.e. x ∈ X. Proof. If T is ergodic then I is the trivial σ-algebra NR consisting of sets of measure 0 and 1. If f ∈ L1 (X, B, µ) then E(f | N ) = f dµ. The result follows from the general version of Birkhoff’s ergodic theorem. 2 §21.4 Appendix: The proof of Birkhoff ’s Ergodic Theorem The proof is something of a tour de force of hard analysis. It is based on the following inequality. Theorem 21.4 (Maximal Inequality) Let (X, B, µ) be a probability space, let T : X → X be a measure-preserving transformation and let f ∈ L1 (X, B, µ). Define f0 = 0 and, for n ≥ 1, fn = f + f ◦ T + · · · + f ◦ T n−1 . For n ≥ 1, set Fn = max fj 0≤j≤n (so that Fn ≥ 0). Then Z f dµ ≥ 0. {x|Fn (x)>0} Proof. Clearly Fn ∈ L1 (X, B, µ). For 0 ≤ j ≤ n, we have Fn ≥ fj , so Fn ◦ T ≥ fj ◦ T . Hence Fn ◦ T + f ≥ fj ◦ T + f = fj+1 and therefore Fn ◦ T (x) + f (x) ≥ max fj (x). 1≤j≤n If Fn (x) > 0 then max fj (x) = max fj (x) = Fn (x), 1≤j≤n 0≤j≤n so we obtain that f ≥ F n − Fn ◦ T 4 MATH41112/61112 Ergodic Theory Lecture 21 on the set A = {x | Fn (x) > 0}. Hence Z Z Z Fn ◦ T dµ Fn dµ − f dµ ≥ A ZA ZA = Fn dµ − Fn ◦ T dµ X A Z Z Fn ◦ T dµ Fn dµ − ≥ X X = 0 where we have used (i) Fn = 0 on X \ A (ii) Fn ◦ T ≥ 0 (iii) µ is T -invariant. 2 Corollary 21.5 If g ∈ L1 (X, B, µ) and if n−1 X 1 Bα = x ∈ X | sup g(T j x) > α n≥1 n j=0 then for all A ∈ B with T −1 A = A we have that Z g dµ ≥ αµ(Bα ∩ A). Bα ∩A Proof. Suppose first that A = X. Let f = g − α, then n−1 ∞ X [ g(T j x) > nα x| Bα = = = n=1 j=0 ∞ [ {x | fn (x) > 0} n=1 ∞ [ {x | Fn (x) > 0} n=1 (since fn (x) > 0 ⇒ Fn (x) > 0 and Fn (x) > 0 ⇒ fj (x) > 0 for some 1 ≤ j ≤ n). Write Cn = {x | Fn (x) > 0} and observe that Cn ⊂ Cn+1 . Thus χCn 5 MATH41112/61112 Ergodic Theory Lecture 21 converges to χBα and so f χCn converges to f χBα , as n → ∞. Furthermore, |f χCn | ≤ |f |. Hence, by the Dominated Convergence Theorem, Z Z Z Z f dµ = f χCn dµ → f χBα dµ = f dµ, as n → ∞. Cn X X Bα Applying the maximal inequality, we have, for all n ≥ 1, Z f dµ ≥ 0. Cn Therefore Z f dµ ≥ 0, Bα i.e., Z g dµ ≥ αµ(Bα ). Bα For the general case, we work with the restriction of T to A, T : A → A, and apply the maximal inequality on this subset to get Z g dµ ≥ αµ(Bα ∩ A), Bα ∩A as required. 2 Proof of Birkhoff ’s Ergodic Theorem. Let n−1 1X f (T j x) f (x) = lim sup n n→∞ ∗ j=0 and n−1 1X f (T j x). f∗ (x) = lim inf n→∞ n j=0 Writing n−1 an (x) = 1X f (T j x), n j=0 observe that n+1 1 an+1 (x) = an (T x) + f (x). n n Taking the lim sup and lim inf as n → ∞ gives us that f ∗ ◦ T = f ∗ and f∗ ◦ T = f ∗ . We have to show (i) f ∗ = f∗ µ-a.e 6 MATH41112/61112 Ergodic Theory Lecture 21 (ii) f ∗ ∈ L1 (X, B, µ) R R (iii) f ∗ dµ = f dµ. We prove (i). For α, β ∈ R, define Eα,β = {x ∈ X | f∗ (x) < β and f ∗ (x) > α}. Note that {x ∈ X | f∗ (x) < f ∗ (x)} = [ Eα,β β<α, α,β∈Q f∗ (a countable union). Thus, to show that = f∗ µ-a.e., it suffices to show that µ(Eα,β ) = 0 whenever β < α. Since f∗ ◦ T = f∗ and f ∗ ◦ T = f ∗ , we see that T −1 Eα,β = Eα,β . If we write n−1 X 1 f (T j x) > α Bα = x ∈ X | sup n≥1 n j=0 then Eα,β ∩ Bα = Eα,β . Applying Corollary 21.5 we have that Z Z f dµ = f dµ Eα,β ∩Bα Eα,β ≥ αµ(Eα,β ∩ Bα ) = αµ(Eα,β ). Replacing f , α and β by −f , −β and −α and using the fact that (−f ) ∗ = −f∗ and (−f )∗ = −f ∗ , we also get Z f dµ ≤ βµ(Eα,β ). Eα,β Therefore αµ(Eα,β ) ≤ βµ(Eα,β ) and since β < α this shows that µ(Eα,β ) = 0. Thus f ∗ = f∗ µ-a.e. and n−1 1X f (T j x) = f ∗ (x) lim n→∞ n j=0 We prove (ii). Let Then gn ≥ 0 and n−1 1 X j gn = f ◦ T . n j=0 Z gn dµ ≤ 7 Z |f | dµ µ-a.e. MATH41112/61112 Ergodic Theory Lecture 21 so we can apply Fatou’s Lemma to conclude that lim n→∞ gn = |f ∗ | is integrable, i.e., that f ∗ ∈ L1 (X, B, µ). We prove (iii). For n ∈ N and k ∈ Z, define k+1 k ∗ n Dk = x ∈ X | ≤ f (x) < . n n For every ε > 0, we have that Dkn ∩ B k −ε = Dkn . n Since T −1 Dkn = Dkn , we can apply Corollary 22.4 again to obtain Z k − ε µ(Dkn ). f dµ ≥ n n Dk Since ε > 0 is arbitrary, we have Z k f dµ ≥ µ(Dkn ). n n Dk Thus Z f ∗ dµ ≤ Dkn ≤ k+1 µ(Dkn ) n Z 1 n µ(Dk ) + f dµ n Dkn (where the first inequality follows from the definition of D kn ). Since [ X= Dkn k∈Z (a disjoint union), summing over k ∈ Z gives Z Z 1 ∗ f dµ µ(X) + f dµ ≤ n X X Z 1 = + f dµ. n X Since this holds for all n ≥ 1, we obtain Z Z f dµ. f ∗ dµ ≤ X X Applying the same argument to −f gives Z Z ∗ (−f ) dµ ≤ −f dµ 8 MATH41112/61112 Ergodic Theory so that Therefore Z ∗ f dµ = Z Z Lecture 21 f∗ dµ ≥ ∗ f dµ = Z Z f dµ. f dµ, as required. Finally, we prove that f ∗ = E(f | I). First note that as f ∗ is T -invariant, it is measurable with respect to I. Moreover, if I is any T -invariant set then Z Z f, dµ = f ∗ dµ. I I Hence f ∗ = E(f | I). 2 9