Let {Xn} be a Markov chain with state space S and transition matrix P = (Pij). Define Pi to be the conditional probability that the Markov chain starts from i, that is, Pi(A) = P(A|X0 = 1) for all events A. Define the return time to i as Ri = min{n ≥ 1 : Xn = i}, and let its p.m.f under Pi be {f(n)}n≥0. That is, we let f(n) = Pi(Ri = n) for all n ≥ 0. Note that f(0) = 0 since Ri is at least 1 by definition.
(a) What does it mean by saying i is recurrent? What does it mean by saying j is transient? Write down both the intuitive meaning and the mathematical definition.
(b) Show that Pi(Ri = 2) = P(2)ii − (Pii)^2, where P(2)ii is the 2-step transition probability from i to itself.
(c) Compute Pi(Ri = 3) in terms of P.