No label defined (Q4940266)

Let \(\{ w_{k},k\geq 1\}\) be a sequence of independent Wiener processes on \([0,T]\) defined on \(\Omega\), where \(\{\Omega,{\mathcal F},P\}\) is a probability space, \({\mathcal F}=\sigma\{w_{k},k\geq 1\}.\) The author considers a diffusion process \(x\) that satisfies the stochastic equation on \([0,T]\) \[ dx(t)=a(x(t),t) dt+\sum_{k=1}^{\infty} b_{k}(x(t),t) dw_{k}(t),\quad x(0)=0. \tag{1} \] The functions \(a: R\times[0,T]\to R\) and \(b=(b_{k})_{k=1}^{\infty}: R\times[0,T]\to l^2\) have three bounded continuous derivatives. Under some conditions equation (1) has a unique strong solution and the process \(x\) has a distribution on the space \(C([0,T])\) with the uniform norm. Let \(\{ f_{ki};k,i\geq q\}\) be a family of functions from \(L_2([0,T])\) and let \(G\) be the \(\sigma\)-field generated by random variables \(\int_0^{T} f_{ki}(t) dw_{k}(t)\), where \(k,i\geq t.\) The author considers the family of random measures \(\{\nu(r,s,t);r\in R;s,t\in [0,T]\}\) which is determined in the following way: \[ \forall f\in C_0(R), \quad \int_{R}f(q) \nu(r,s,t) (dq):= E(f(x(r,s,t))/G), \] where \(\{ x(\cdot,s,t);0\leq s\leq t\leq T\}\) is the continuous stochastic flow associated with (1). The family \(\{\nu(r,s,t);r\in R;s,t\in [0,T]\}\) does not satisfy the Chapman-Kolmogorov equation in general. The author proves that under some assumptions the integrals \(\int_{R}f(q)\nu(r,s,t) (dq)\) satisfy some anticipating stochastic equation.