The class of tenable zero-balanced Pólya urns with an initially dominant subset of colors (Q654467)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The class of tenable zero-balanced Pólya urns with an initially dominant subset of colors
scientific article

    Statements

    The class of tenable zero-balanced Pólya urns with an initially dominant subset of colors (English)
    0 references
    0 references
    0 references
    28 December 2011
    0 references
    This paper considers a generalised Pólya urn scheme in which there is a collection of balls of up to \(k\) different colours, and at each stage a ball is selected uniformly at random and if it has colour \(r\), then it is put back together with \(A_{r,s}\) balls of colour \(s=1,\dots, k\). It is assumed that \(\sum_{s=1}^{k} A_{r,s}=0\), whereby the number of balls in the urn remains the same at each stage. Moreover, it is assumed that the scheme satisfies a number of conditions which ensure that the process can continue indefinitely no matter what ``path'' has been followed and that all colours are essential, and we cannot treat any two colours as one. Assuming that \(n\) is the total number of balls and \(n \rightarrow \infty\), let the process run for \(j_n\) stages. The assumption is that initially a subset of colours is dominant in the sense that, for each one of these colours, the initial fraction of balls having this colour is positive, whereas the balls having one of the remaining colours occupy a sublinear (in \(n\)) portion of \(n\). The authors use a martingale central limit theorem to deduce the joint asymptotic distribution (as \(n\) and \(j_n\) grow) of any given collection of colours. They also do the same when one colour is initially dominant.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Pólya urns
    0 references
    central limit theorem
    0 references
    martingales
    0 references
    0 references