Paper 1, Section II, J

Applied Probability | Part II, 2013

Let (Xt,t0)\left(X_{t}, t \geqslant 0\right) be a Markov chain on {0,1,}\{0,1, \ldots\} with QQ-matrix given by

qn,n+1=λnqn,0=λnεn(n>0)qn,m=0 if m{0,n,n+1}\begin{aligned} q_{n, n+1} &=\lambda_{n} \\ q_{n, 0} &=\lambda_{n} \varepsilon_{n} \quad(n>0) \\ q_{n, m} &=0 \quad \text { if } m \notin\{0, n, n+1\} \end{aligned}

where εn,λn>0\varepsilon_{n}, \lambda_{n}>0.

(i) Show that XX is transient if and only if nεn<\sum_{n} \varepsilon_{n}<\infty. [You may assume without proof that x(1δ)log(1+x)xx(1-\delta) \leqslant \log (1+x) \leqslant x for all δ>0\delta>0 and all sufficiently small positive xx.]

(ii) Assume that nεn<\sum_{n} \varepsilon_{n}<\infty. Find a necessary and sufficient condition for XX to be almost surely explosive. [You may assume without proof standard results about pure birth processes, provided that they are stated clearly.]

(iii) Find a stationary measure for XX. For the case λn=λ\lambda_{n}=\lambda and εn=α/(n+1)(λ,α>0)\varepsilon_{n}=\alpha /(n+1)(\lambda, \alpha>0), show that XX is positive recurrent if and only if α>1\alpha>1.

Typos? Please submit corrections to this page on GitHub.