Perturbed Markov chains (Q4408538)

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 1941397
Language Label Description Also known as
default for all languages
No label defined
    English
    Perturbed Markov chains
    scientific article; zbMATH DE number 1941397

      Statements

      Perturbed Markov chains (English)
      0 references
      0 references
      0 references
      7 December 2003
      0 references
      Markov chains
      0 references
      stationary distribution
      0 references
      exit distribution
      0 references
      conductance
      0 references
      sensitivity analysis
      0 references
      perturbation theory
      0 references
      stability of a Markov chain
      0 references
      This paper contains irreducible time-homogeneous Markov chains with a finite state space in discrete time. There are interesting the effects of perturbation of the transition matrix on the stationary distribution and other statistical quantities [see \textit{P. J. Schweitzer}, J. Appl. Probab. 5, 401-413 (1968; Zbl 0196.19803)]. Let \(S\) be a finite set and \(q=(q(t)s)_{s,t\in S}\) an irreducible transition matrix. Let \(\widetilde q=(\widetilde q(t)s)_{s,t\in S}\) be another transition matrix. Denote by \(\mu=(\mu_s)_{s\in S}\) and \(\widehat \mu=(\widehat \mu_s)_{s\in S}\) the stationary distributions that correspond to \(q\) and \(\widehat q\), respectively. Schweitzer (loc. cit.) estimated \(|\mu-\widehat\mu |_\infty\) using the fundamental matrix of \textit{J. G. Kemeny} and \textit{J. L. Snell} \ [``Finite Markov chains'' (1960; Zbl 0089.13704)]. Important literature has followed up Schweitzer (see the authors' references). \textit{C. A. O'Cineide} [Numer. Math. 65, 109-120 (1993; Zbl 0804.65145)] studied the effects relative to the perturbations on the stationary distribution. Section 2 of the paper contains the statements of the main results. Also the authors study an example that illustrates the advantage of the closeness relation and compare the bound they derive to existing bounds. Section 3 recalls standard formulae, and states a few elementary properties of the graphs. Section 4 treats the proof of the main result. The last section deals with a variation of the main result.
      0 references

      Identifiers