Strong law of large numbers and mixing for the invariant distributions of measure-valued diffusions. (Q2574566)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Strong law of large numbers and mixing for the invariant distributions of measure-valued diffusions.
scientific article

    Statements

    Strong law of large numbers and mixing for the invariant distributions of measure-valued diffusions. (English)
    0 references
    0 references
    29 November 2005
    0 references
    Let us denote by \(\mathbf M\) the space of all locally finite Borel measures on \(\mathbb R^ {d}\), endowed with the vague topology, and consider diffusion processes on \(\mathbf M\) whose underlying motion is a conservative reversible diffusion on \(\mathbb R^ {d}\) with a generator \(L_ 0 = \frac 12 \nabla \cdot a\nabla + (a\nabla Q)\cdot \nabla \) and whose branching mechanism is of the form \(\beta (x)z - \alpha (x)z^ 2\). It is supposed that the matrix \(a\) is positive definite, \(a_ {ij}\), \(Q\) are \(C^ 1\)-smooth functions with Hölder continuous derivatives, \(\alpha , \beta \) are Hölder, \(\alpha >0\), \(\beta \) bounded from above. The author studied recently [Ann.\ Probab.\ 29, 1476--1514 (2001; Zbl 1108.60312)] the problem of existence of invariant measures for such \(\mathbf M\)-valued diffusions. Sufficient conditions are found for these invariant measures to satisfy the strong law of large numbers with respect to functions \(f\) is a suitable class or to be mixing. We say that a probability measure \(\nu \) on \(\mathbf M\) satisfies the strong law of large numbers with respect to a nonnegative locally bounded function \(f\) provided that the mean measure \(\pi _ \nu \), \(\pi _ \nu (B) = \int _ {\mathbf M} \eta (B)\,d\nu (\eta )\) for Borel sets \(B\subseteq \mathbb R^ {d}\), is locally finite, \(\langle f,\pi _ \nu \rangle =\infty \) and \(\langle f_ {n},\eta \rangle /\langle f_ {n},\pi _ \nu \rangle \to 1\) as \(n\to \infty \) for \(\nu \)-almost all \(\eta \in \mathbf M\), whenever \(\{f_ {n}\}\) is an increasing sequence of compactly supported functions that converge to \(f\). The measure \(\nu \) is called mixing with respect to sequences \(\{A_ {n}\}\), \(\{B_ {n}\}\) of sets, if \(\int _ {\mathbf M} f(\eta (A_ {n}))g(\eta (B_ {n}))\,d\nu (\eta ) - \int _ {\mathbf M} f(\eta (A_ {n}))\,d\nu (\eta ) \int _ {\mathbf M} g(\eta (B_ {n}))\,d\nu (\eta )\) converges to 0 as \(n\to \infty \) for all \(C^ 1\)-smooth bounded functions \(f,g\) on \(\mathbb R_ {+}\). It is shown that the obtained results are applicable to many interesting nontrivial examples.
    0 references
    0 references
    0 references
    invariant measures
    0 references