Running list of ergodic theory exercises : 1. Show that the collection of all finite disjoint unions of elements of a semi-algebra is an algebra. 2. Let U = (u1 , . . . , un ) ∈ Rn (viewed as a function on {1, . . . , n}). Let β ∈ R. Let µβ be the probability vector (p1 , . . . , pn ) defined by pi = e−βui Z(β) where Z is the normalization factor: Z(β) = X e−βuj . j Show that dEµβ (U ) = −Varianceµβ (U ). dβ P 3. Show that the collection of functions Hn (p1 , . . . , pn ) := ni=1 −pi log2 pi (defined on all probability vectors (p1 , . . . , pn )), is uniquely determined by the following axioms: (a) each Hn is a symmetric function of (p1 , . . . , pn ). (b) H2 (1/2, 1/2) = 1 (c) H2 (p, 1 − p) is a continuous function of p. (d) Hn (p1 , . . . , pn ) = Hn−1 (p1 + p2 , p3 , . . . , pn ) + (p1 + p2 )H2 (p1 /(p1 + p2 ), p2 /(p1 + p2 )) 4. Complete the proof, from class, that the following are equivalent. 1a T is ergodic 1b if µ(A) > 0, a.e. x ∈ M visits A infinitely often, i.e., −k −n µ(∩∞ (∪∞ A)) = 1. n=1 T k=1 T 2a if µ(A), µ(B) > 0, then for some n > 0, µ(T −n (A) ∩ B) > 0. 2b if µ(A), µ(B) > 0, then for infinitely many n > 0, µ(T −n (A) ∩ B) > 0. 3a (the usual definition) If T −1 (A) = A, then µ(A) = 0 or 1. 3b If T −1 (A) ⊆ A, then µ(A) = 0 or 1. 3c If T −1 (A) ⊇ A, then µ(A) = 0 or 1. 3d If µ(T −1 (A) \ A) = 0, then µ(A) = 0 or 1. 3e If µ(A \ T −1 (A)) = 0, then µ(A) = 0 or 1. 3f If µ(T −1 (A)∆A) = 0, then µ(A) = 0 or 1. 4a If f ∈ Lp and f ◦ T = f , then f is constant a.e. (here, 0 ≤ p ≤ ∞) 4b If f ∈ Lp and f ◦ T = f a.e., then f is constant a.e. (here, 0 ≤ p ≤ ∞) 1 5. Let T be an MPT and I denote the collection of measurable sets that are invariant mod measure zero, i.e., µ(A∆T −1 (A)) = 0. Show that I is a σ-algebra. 6. Let X and Y be stationary processes. Let T + and S + be the MPT’s corresponding to the one-sided versions of X and Y . Let T and S be the MPT’s corresponding to the two-sided versions of X and Y . Show that if T + and S + are isomorphic, then T and S are isomorphic. 7. Let T + and S + be the MPT’s corresponding to the one-sided versions of iid(p) and iid(q), where p and q are strictly positive probability vectors, not necessarily of the same length. Show that T + and S + are isomorphic iff p and q are the same, up to a reordering (in particular, they have the same length). 8. Suppose that P is a stochastic matrix and π is a strictly positive probability vector π s.t. πP = π. Define the relation on the vertices of the graph G(P ) by i ∼ j if there is a path from i to j. Show that this relation is an equivalence relation (the equivalence classes of this relation are called irreducible components). 9. Consider the MPT defined by a one-sided stationary Markov chain, given by P and π as above. Let Ai = {x ∈ F Z+ : x0 = i}. Give an interpretation to the limiting function χ∗Ai (from the ergodic theorem) in terms of the irreducible components of P . 10. Show that mixing is an isomorphism invariant and more generally that if there is a measure preserving homomorphism from T onto S and T is mixing, then so is S. Find an example of MPT’s T and S and a measure preserving homomorphism from T onto S such that S is mixing but T is not mixing. 11. Let T be an invertible MPT. Show that T is ergodic (resp., mixing) iff T −1 is ergodic (resp., mixing). 12. Let T be an MPT and B a semi-algebra which generates the underlying σ-algebra. Show that if µ(T −n (A) ∩ B) → µ(A)µ(B). for all A, B ∈ B, then T is mixing. 13. Complete the proof that a stochastic matrix is primitive iff it is irreducible and aperiodic. 14. Complete the geometric proof (using convexity ideas, given in class) that if P is a primitive stochastic matrix with stationary vector π, then for all i, j (P n )ij → πj 2