Fubini theorem w.r.t. stochastic measure on product measurable space (Q1891420): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
Removed claims |
||
Property / author | |||
Property / author: Pei-de Chen / rank | |||
Property / reviewed by | |||
Property / reviewed by: Aurel Spătaru / rank | |||
Revision as of 03:41, 12 February 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Fubini theorem w.r.t. stochastic measure on product measurable space |
scientific article |
Statements
Fubini theorem w.r.t. stochastic measure on product measurable space (English)
0 references
27 September 1995
0 references
Let \((X, {\mathcal X})\) and \((Y, {\mathcal Y})\) be measurable spaces, let \((\Omega, {\mathcal F}, P)\) be a complete probability space, and let \(L = L(\Omega, {\mathcal F},P)\) denote the set of all nonnegative a.s. finite measurable functions from \(\Omega\) into \(R\). A \(\sigma\)-additive set function \(\mu : {\mathcal X} \to L\) is said to be a stochastic measure. For two stochastic measures \(\mu : {\mathcal X} \to L\) and \(\nu : {\mathcal Y} \to L\), the authors assert the existence of a stochastic measure \(\mu \otimes \nu : {\mathcal X} \otimes {\mathcal Y} \to L\) such that \((\mu \otimes \nu) (E \times F) = \mu (E) \nu (F)\), \(E \in {\mathcal X}\), \(F \in {\mathcal Y}\). Let \(f : X \times Y \to R\) be a measurable function such that \(\int_{X \times Y} f^ +d(\mu \otimes \nu)\), \(\int_{X \times Y} f^ -d (\mu \otimes \nu) \in L\). The authors prove that \[ \int_{X \times Y} fd (\mu \otimes \nu) = \int_ X \Bigl( \int_ Yfd \overline \nu \Bigr) d \widehat \mu = \int_ Y \Bigl( \int_ X fd \overline \mu \Bigr) d \widehat \nu, \] where \(\widehat \mu : {\mathcal X} \otimes {\mathcal F} \to L\) is a stochastic measure determined by \(\widehat \mu (E \times A) = \mu (E) \mathbf{1}_ A\), \(E \in {\mathcal X}\), \(A \in {\mathcal F}\), \(\overline \mu : {\mathcal X} \otimes {\mathcal Y} \to L (Y \times \Omega, {\mathcal Y} \otimes {\mathcal F}, Q)\) (here \(Q\) is a suitable probability measure on \({\mathcal Y} \otimes {\mathcal F})\) is a stochastic measure determined by \(\overline \mu (E \times F) = \mu (E) \mathbf{1}_ F\); \(E \in {\mathcal X}\), \(F \in {\mathcal Y}\), etc. They also present other extensions of the Fubini theorem, and discuss versions of the Radon-Nikodym theorem and of the Lebesgue decomposition theorem for pairs of bounded signed stochastic measures. The reviewer feels that some parts of this theory of integration with respect to stochastic measures is useless. This feeling is supported by the authors attempt to illustrate applicability of their results. Thus they ``shoot with the cannon at sparrows'' to show that \(\int_{[0,1]^ 2} f(x,y)d \lambda (x,y) = \int^ 1_ 0 f(x,x)dx\), where \(\lambda\) is the Lebesgue-Stieltjes measure on \([0,1]^ 2\) generated by \(F(x,y) = x \wedge y\), \(x,y \in [0,1]\). The paper is full of grammatical errors.
0 references
Radon-Nikodym theorem
0 references
Lebesgue decomposition theorem
0 references
bounded signed stochastic measures
0 references