scientific article; zbMATH DE number 4026462

From MaRDI portal
Publication:3768096





[For the entire collection see Zbl 0626.00026.] The purpose of this article is to investigate the continuity properties of the filtering measures for systems governed by stochastic differential equations. Let X(t) and Y(t) be given by \[ (1)\quad dX(t)=\sigma (t,X(t))dZ(t)+b(t,X(t))dt,\quad X(0)=X_ 0, \] \[ (2)\quad dY(t)=dW(t)+h(t,X(t))dt,\quad Y(0)=0, \] where Z(t) and W(t) are mutually independent (multi-dimensional) Wiener processes, X(t) is the signal process and Y(t) is the observable process. The filtering measure \(\Pi\) (\(\cdot | y)\) is the conditional probability measure of X(s), \(0\leq s\leq t\), under the condition that \(Y(s)=y(s)\), \(0\leq s\leq t\). The measure \(\Pi\) (\(\cdot | y)\) depends on the coefficients of the system (1) and (2), i.e. \(\sigma\) (t), b(t) and h(t), and on the observed path y(s), \(0\leq s\leq t\). The stability of the filtering system (1) and (2) is, by definition, the continuity of the measure \(\Pi\) (\(\cdot | y)\) with respect to these coefficients and the observed path. The main result of this paper is stated in Theorem 1 of {\S} 3, which gives us certain sufficient conditions for the stability of the filtering systems. The advantage of this result is that this theorem can be applied to systems whose coefficients are not necessarily bounded. For the proof of this result, the so-called unnormalized filtering measure method is used.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3768096)