<p>The (<i>k</i>,&#xa0;<i>r</i>)-Quicksort on the fly is a sorting algorithm, which provides successively first the smallest, then the second smallest, and so on of a given set of size <i>n</i>. Let <i>X</i>(<i>n</i>,&#xa0;<i>l</i>) be the number of comparisons (respectively the time) up to the <i>l</i>th smallest shown. The correctly normalized process <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(Y_n\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>Y</mi> <mi>n</mi> </msub> </math></EquationSource> </InlineEquation> provides in the limit a cadlag process <i>Y</i>. The process <i>Y</i> is characterized as a stochastic fixed point. For suitable versions, the convergence <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(Y_n\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>Y</mi> <mi>n</mi> </msub> </math></EquationSource> </InlineEquation> to <i>Y</i> is in Skorokhod metric a.e. On the way, we also provide the average number of comparisons for the (<i>k</i>,&#xa0;<i>r</i>)-Quicksort algorithm.</p><p>The method of proof might be interesting in itself. It uses the contraction method and weighted branching processes.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

On the (kr)-Quicksort on the fly process

  • Mehri Javanian,
  • Uwe Roesler

摘要

The (kr)-Quicksort on the fly is a sorting algorithm, which provides successively first the smallest, then the second smallest, and so on of a given set of size n. Let X(nl) be the number of comparisons (respectively the time) up to the lth smallest shown. The correctly normalized process \(Y_n\) Y n provides in the limit a cadlag process Y. The process Y is characterized as a stochastic fixed point. For suitable versions, the convergence \(Y_n\) Y n to Y is in Skorokhod metric a.e. On the way, we also provide the average number of comparisons for the (kr)-Quicksort algorithm.

The method of proof might be interesting in itself. It uses the contraction method and weighted branching processes.