Normal approximation for coverage models over binomial point processes (Q968780)

From MaRDI portal
Revision as of 18:53, 2 July 2024 by ReferenceBot (talk | contribs) (‎Changed an Item)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
Normal approximation for coverage models over binomial point processes
scientific article

    Statements

    Normal approximation for coverage models over binomial point processes (English)
    0 references
    0 references
    0 references
    6 May 2010
    0 references
    Given a collection of \(n\) independent uniformly distributed random points in a \(d\)-dimensional cube of volume \(n\) (the so-called binomial point process), let \(V\) denote the random total volume of the union of interpenetrating balls of fixed radius \(\rho\) centered at these points, and let \(S\) denote the number of balls of radius \(\rho/2\) (centered at the same set of points) which are singletons, that is, do not overlap any other such ball. These variables are fundamental topics of interest in the stochastic geometry of coverage processes and random geometric graphs [[1]: \textit{M. D. Penrose}, Random geometric graphs. Oxford Studies in Probability. 5. Oxford: Oxford University Press (2003; Zbl 1029.60007)]. As \(n\to \infty\) with \(\rho\) fixed (the so-called thermodynamic limit), both \(V\) and \(S\) are known to satisfy a central limit theorem (CLT) (see [1]). In the present paper, an associated Berry-Esseen type result is provided, that is, the authors show under periodic boundary conditions that the cumulative distribution functions converge to that of the normal at the same \(O(n^{-1/2})\) rate as for a sum of \(n\) independent identically distributed variables, and provide bounds on the quantity of the normal approximation for finite \(n\). Were the authors to consider instead a Poisson-distributed number of points, that is, a Poisson point process instead of binomial one, both of the variables of interest, and thereby Berry-Esseen type bounds could be obtained by known methods. But with a nonrandom number of points, the local dependence is lost and the de-Poissonization arguments in [1] and [[2]: \textit{M. D. Penrose} and {J. E. Yukich}, Ann. Appl. Probab. 11, No. 4, 1005--1041 (2001; Zbl 1044.60016)] do not provide error bounds for the de-Poissonized CLTs. The variables \(V\) and \(S\) are just two of a large class of variables of interest that can be expressed as a sum, over the \(n\) points, of terms that depend only on the configuration of nearby points in some sense. General CLTs have been developed for such variables, see [2] and [[3]: \textit{M. D. Penrose}, Electron. J. Probab. 12, 989--1035 (2007; Zbl 1153.60015)] and general Berry-Esseen type results are available in the Poissonized setting, but it remains open to provide a generally applicable Berry-Esseen type result for such sums when \(n\) is nonrandom, see [[4]: \textit{S. Chatterjee}, Ann. Probab. 36, No. 4, 1584--1610 (2008; Zbl 1159.62009)]. However, there seems to be good prospects of adapting the approach of the present paper (which is new in the geometrical setting) to a wider class of geometrical sums. The authors give error bounds which demonstrate optimal rates of convergence in the CLT for \(V\) and \(S\), for germ-grain models with fixed grain radius over a binomial point process of \(n\) points in a toroidal spatial region of volume \(n\).
    0 references
    coverage process
    0 references
    Berry-Esseen theorem
    0 references
    size biased coupling
    0 references
    Stein's method
    0 references

    Identifiers

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