Estimation of the evolution of a random set (Q2362415)

The \(n\)-vector Ito differential equation \( dX_t = b(t, X_t) dt + \sigma (t, X_t) dW_t\), \(t\in [0,T]\), with \(W_t\) is the standard Wiener \(m\)-dimensional process is considered. Assume that the initial random vector \(\xi\) is unknown and all that is given is an inclusion \(\xi (\omega) \in S(\omega)\) a.s. where \(S\) is a random closed set (RCS). The initial RCS \(S_0\) is transformed into an RCS \(S_t\) by the stochastic differential equation. An estimation problem for the random set is considered. The Markov property is proven for it. A trick is used: the random initial set of the differential equation is approximated by a finite set on an integer multidimensional grid, and the differential equation is approximated by a multistep Markov chain. Numerical examples are given.