Identification of spatially varying parameters in distributed parameter systems by discrete regularization (Q1820334)

The authors consider the problem of identifying spatially varying parameters in parabolic partial differential equations from a finite set of noisy measurements. To deal with the inherent ill-posedness, they apply a Tikhonov regularization to stabilize the problem. For the numerical treatment, the authors propose a discretized minimization of the smoothing functional by using finite-dimensional approximations in Sobolev espaces.