Robustness and convergence of approximations to nonlinear filters for jump-diffusions (Q1375892): Difference between revisions

The author considers the nonlinear filtering problem for a jump-diffusion reflected process \(x(\cdot)\) of the form \[ x(t)=x+ \int^t_0 b\bigl(x(s) \bigr)ds +\int^t_0 \sigma \bigl(x(s) \bigr)dw(s)+J(t)+z(t), \] where \(J(\cdot)\) is a jump term and \(z(\cdot)\) is a reflection term. The observation process \(y(\cdot)\) is defined by \[ dy(t)= g\bigl(x(t) \bigr)dt+dw_0 (t). \] The process \(x(\cdot)\) is approximated in the sense of weak convergence by a Markov chain for which the filter can be computed. The observational data which is used is the actual physical data. Theorems of convergence and robustness are proved. The Monte Carlo methods are considered as well. The fundamental approach was introduced by the author in [\textit{H. J. Kushner}, Probability Methods for Approximations in Stochastic Control and for Elliptic Equations, Academic Press, New York (1977; Zbl 0547.93076) and Stochastics 3, 75-83 (1979; Zbl 0421.60054)].