No label defined (Q3322960)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 3854136
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 3854136

      Statements

      0 references
      1983
      0 references
      Lévi-Prohorov metric
      0 references
      uniform-in-time continuity
      0 references
      ideas of coupling
      0 references
      convergence rates
      0 references
      [For the entire collection see Zbl 0522.00010.] Let \(\{\xi_ t\}_{t\geq 0}\) be a sequence of i.i.d. r.v.'s taking values from a complete separable metric space (CSMS) \((U,\rho)\). Let (V,h) be a CSMS and F: \(U\times V\to V\). Let \(Y_ 0\) be an arbitrary r.v. with values from V and (1) \(Y_{t+1}=F(Y_ t,\xi_ t)\), \(t=0,1,...\). Then the sequence \(\{Y_ t\}_{t\geq 0}\) is a Markov chain. Let \(\{\Gamma^ V_ i\}\), \(\{\Gamma^ U_ i\}\) be some compacts in (V,h), \((U,\rho)\) correspondingly meeting the relations: \(\Gamma^ V_ 1\subset \Gamma^ V_ 2\subset...\), \(\Gamma^ U_ 1\subset \Gamma^ U_ 2\subset...\), \(\cup_{i}\Gamma^ V_ i=V\), \(\cup_{i}\Gamma^ U_ i=U\). Let also \(\pi\) be a Lévi-Prohorov metric. In order to give an impression about the results we display one of the theorems. Theorem. Let the Markov chains \(\{Y_ t\}\) and \(\{Y_ t^{(n)}\}\) be of the form (1) (but with different \((Y_ 0,\{\xi_ t\},F)\) and consider \((Y_ 0^{(n)},\{\xi_ t^{(n)}\},F^{(n)}))\). Then \(Y_ t^{(n)}\to^{w}Y_ t\) for any \(t\geq 0\) if \(Y_ 0^{(n)}\to^{w}Y_ 0\), \(\xi_ 0^{(n)}\to^{w}\xi_ 0\), \(\sup_{y\in \Gamma^ V_ N,\xi \in \Gamma^ U_ L}h(F^{(n)}(y,\xi),\quad F(y,\xi))\to_{n\to \infty}0 \forall N,L\) and the mapping F is continuous. Imposing more restrictions (concerning, e.g. the recurrence properties of the Markov chains under comparison) the author obtains conditions implying the uniform-in-time continuity when \(\sup_{t}\pi(Y_ t^{(n)},Y_ t)\to_{n\to \infty}0\). Similar results are obtained in terms of other metrics. Because of the constructive character of the proofs (which used some ideas of coupling) the author receives the estimates of convergence rates in the above statements. He constructs also a finite approximation for the above general Markov chains. The applications are from queueing.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references