Generalized inverse extrapolation of stochastic processes by an aggregate of continuous discrete observations with memory (Q1386990): Difference between revisions

In the case of continuous observations the considered system is of the form (in the sense of Ito) \[ dx_t= f(t,x_t)dt +\Phi_1 (t,x_t)dw_t, \] \[ dz_t =h(t,x_t,x_{\tau_1}, \dots, x_{\tau_N }, z_t)dt +\Phi_2 (t,z_t) dv_t, \] where \(x_t\) is an \(n\)-dimensional unobserved process, \(z_t\) is an \(l\)-dimensional observed process, \(0<\tau_N <\tau_{N-1} <\cdots <\tau_1 \leq t\). The authors derive the equations of nonlinear extrapolation for mean square optimal estimations \(\mu(s_k,t)\) of the process values \(x_{s_k}\), \(k=1, \dots,K\), where the fixed time moments \(s_1, \dots, s_K\) are such that \(t<s_1 <s_2< \cdots <s_K\). The case of continuous together with discrete observations is considered as well.