Embedding a stochastic difference equation into a continuous-time process

Infinitely divisible distributions; stable distributions (60E07) Stochastic analysis (60H99) Continuous-time Markov processes on general state spaces (60J25)

Review text

Let \(\{Y_ n\}\), \(n\geq 0\), be a sequence of random variables satisfying \[ (1)\quad Y_ n=A_ nY_{n-1}+B_ n,\quad n\geq 1, \] subject to the condition (2): \(\{A_ n,B_ n\}\) are i.i.d. and \(A_ 1>0\) with probability 1. A continuous-time process \(\{X_ t\}\), \(t\geq 0\), is said to be divisible if the sequence \(\{X_{nh}\}\), \(n\geq 0\), satisfies relations of the type (1) and (2) for all \(h>0\). Properties of such a process are studied. A representation of \(\{X_ t\}\) in terms of a two-dimensional process with stationary and independent increments is given. Further, under very mild conditions on \((A_ 1,B_ 1)\), it is shown that the stationary initial distribution of the process exists. In such a case, one can embed \(\{Y_ n\}\) in the process \(\{X_ t\}\).

Recommendations

Cites work

Cited in

(14)

This page was built for publication: Embedding a stochastic difference equation into a continuous-time process

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1822836)