No label defined (Q4487683)

Let \(X\) be a set and \({\mathcal B}\) be a \(\sigma\)-algebra of subsets of \(X\). The random measure \(\mu\) is a set of random variables \(\{\mu(A),\;A\in{\mathcal B}\}\) such that \(\mu(A\cup B)=\mu(A)+\mu(B)\) a.s. for \(A\cap B=\emptyset\) and \(\mu(A_n)\overset {P}{} 0\) as \(A_n\downarrow\emptyset\). In the monograph the random measures in such general sense and stochastic integrals with respect to such measures are considered. In the beginning the integral of real functions is defined and later it is generalized to some class of random functions. Limit theorems for such integrals and stochastic integral equations are studied. In the short Chapter 0 the main notations and definitions are introduced and some known statements which are used in the book are given. In Chapter 1 a construction of the integral of a real function with respect to a random measure is given. The integral is defined as the limit of integrals of simple functions and the construction is based on results by M. Talagrand and others on boundedness of \(L_0\)-valued measure. In particular, for any \(\mu\) all real bounded measurable functions are integrable. Limit properties are investigated for integrals defined on a sequence of functions, on a sequence of random measures and on a sequence of sets of an integration. The notion of uniform integrability for the constructed integrals is introduced and used to prove some theorems. In Chapter 2 the \(\sigma\)-finite random functions of sets as a generalization of the random measures are investigated. The integral of a real function is introduced and limit theorems are proved. In Chapter 3 some special cases of measures and integrals are investigated. Random measures which are continuous on an empty set a.s., as well as convergence a.s. of integrals with respect to such measures are considered. For the random measures being at each fixed \(\omega\) real alternating measures the author obtains an analogue of the Radon-Nikodym theorem. The integral of a random function with respect to a real measure as the limit of integrals of simple functions is defined as well as an analogue of the Lebesgue theorem on the limit of a sequence of integrals is proved. In Chapter 4 the integral of the random function \(f(x,\omega)=\sum_{k=1}^{\infty}\xi_{k}(\omega)f_{k}(x),\) where \(\xi_{k}\) are random variables and \(f_{k}: X\to{\mathbb{R}}\) are measurable functions with \(|f_{k}(x)|\leq 1\), is defined as \(\int_{A}f d\mu =\sum_{k=1}^{\infty}\xi_{k}\int_{A}f_{k} d\mu,\) as soon as the series converge in probability unconditionally. This is a generalization of the symmetric stochastic integral. For this integral limit theorems are obtained and stochastic integral equations with respect to the unknown random measure are considered. In Chapter 5 another definition of the integral of a random function on \([0, 1]\) is given. In some sense this is a generalization of Itô's integral with respect to the Wiener process. The integral is defined at finite unions of binary half-intervals. The limit theorems for this integral are proved and stochastic integral equations with respect to the unknown random measure are investigated. Note that conditions, under which the proposed theorems are proved, are very restrictive. These conditions are satisfied, for example, by a.s. Lipschitz random functions.