Identification of transmissivity coefficients by mollification techniques. II: \(n\)-dimensional elliptic problems (Q1903761): Difference between revisions

As a continuation of Part I [ibid. 25, No. 8, 59-79 (1993; Zbl 0774.65089)] the case of identifying transmissivity coefficients \(a(x)\) in \(n\)-dimensional elliptic problems \[ \text{div} (a(x) \nabla u(x)) = f(x), \quad x \in \Omega \subset \mathbb{R}^n \] with appropriate boundary conditions is considered. A list of arguments for the ill-posedness of these inverse problems is given that yields reasons for the use of mollification methods. Since the multidimensional case is under consideration, mollified versions of partial derivatives are discussed in detail. The author combines mollification techniques and finite differences and gets a numerical method which seems to be easy to implement in rectangular domains. For the two-dimensional case plots of numerical results are presented for the identification of transmissivity coefficients.