Convergence rate for the distributions of <i>GI/M/</i>1/<i>n</i> and <i>M/GI/</i>1/<i>n</i> as <i>n</i> tends to infinity (Q4376523)

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 1117516
Language Label Description Also known as
default for all languages
No label defined
    English
    Convergence rate for the distributions of <i>GI/M/</i>1/<i>n</i> and <i>M/GI/</i>1/<i>n</i> as <i>n</i> tends to infinity
    scientific article; zbMATH DE number 1117516

      Statements

      Convergence rate for the distributions of <i>GI/M/</i>1/<i>n</i> and <i>M/GI/</i>1/<i>n</i> as <i>n</i> tends to infinity (English)
      0 references
      0 references
      1 November 1998
      0 references
      embedded Markov chain
      0 references
      Lindley process
      0 references
      coupling
      0 references
      convergence rate
      0 references
      queueing system
      0 references
      This paper discusses the convergence rate of the limiting distribution \({\mathbf q}_n =(q_n(j))\) and the arrival stationary distribution \(\pi_n= (\pi_n(j))\) of the GI/M/1/\(n\) queueing system to the corresponding distributions \({\mathbf q} =(q(j))\) and \(\pi= (\pi(j))\) of the GI/M/1 system as the capacity size \(n\) goes to infinity. More specifically, the authors identify a number \(R\) such that \[ \lim_{n\to\infty} r^n\| q_n- q\|=0 \text{ if }r>R, \qquad \limsup_{n\to\infty} r^n\| q_n-q \|= +\infty \text{ if } r>R \] where \(\| q_n-q\|= \sum_{j \geq 0} | q_n(j) -q(j) |\). A similar result is established for M/G/1/\(n\) and M/G/1 queueing systems under more stringent assumptions.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references