A Lévy type martingale convergence theorem for random sets with unbounded values (Q1970311)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A Lévy type martingale convergence theorem for random sets with unbounded values
scientific article

    Statements

    A Lévy type martingale convergence theorem for random sets with unbounded values (English)
    0 references
    0 references
    0 references
    12 December 2000
    0 references
    The aim of the paper is to extend the Lévy martingale convergence theorem (LMCT) to the random sets \(F\) whose values are members of \(\mathcal{F}(X)\), the space of nonempty closed sets of a Banach space \(X\). Such multivalued versions of the LMCT are proved for random sets with closed values. More precisely, the authors prove that the sequence \((\mathbb{E}(F\mid \mathcal{B}_n))\) converges to \((\mathbb{E}(F\mid \mathcal{B}_\infty))\) in the sense of Painlevé-Kuratowski (PK). The PK convergence when considered for sequences of unbounded sets, gives rise to a specific functional convergence often called ``epi-convergence'' or ``\(\Gamma\)-convergence'' that is considered in the paper in a stochastic context. The infinite-dimensional extensions such as the Mosco convergence, the Wijsman topology or the slice topology are considered too. Even in finite-dimensional case the results are new or subsume the existing ones. The method relies on countability supportness hypotheses which are shown to hold when the values of the random set \(\mathbb{E}(F\mid \mathcal{B}_\infty)\) do not contain any line. Conditions involving barrier and asymptotic cones are demonstrated to be necessary.
    0 references
    measurable multifunctions
    0 references
    random sets
    0 references
    multivalued conditional expectations
    0 references
    set convergence
    0 references
    epi-convergence
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references