Parameter identification for stochastic diffusion equations with unknown boundary conditions (Q1099124)

Stochastic diffusion equations with unknown boundary conditions are encountered in practice whenever the boundary is observed only through noisy measurements. Parameter identification for such systems is studied in the present paper. The boundary condition is treated as an unknown parameter in a function space as an intermediate step in solving the identification problem, which is then achieved in two steps. First, the maximum a posteriori smoothed estimate of the boundary data is derived and, then, the maximum likelihood estimate of the unknown parameters is obtained. Finally, the proposed method is illustrated by means of an example.