Error estimates for distributed parameter identification in parabolic problems with output least squares and Crank-Nicolson method (Q1265628)

The paper deals with the identification problem of the heat conductivity coefficient of a linear heat equation in two space variables supposing that the distributed observation of the temperature field is available. This inverse problem is solved by transforming it to a minimization problem by means of the least squares method. Also an approximate method is suggested. To be able to estimate the error of this approximation the author has to investigate the error bounds of the approximate solution of a direct parabolic problem when the discretization in space variables is constructed by the finite element method and the time discretization is performed via the Crank-Nicolson scheme. Among others, he finds completely new estimates of the time derivative of the approximate solution.