A domain decomposition method for determining the diffusion coefficient of a two-dimensional linear diffusion equation in the time domain (Q1339341)

A domain decomposition method (DD) for determining the coefficient of a two-dimensional linear diffusion equation in the time domain is considered. The DD method is one of the most important ways of studying the numerical solutions of partial differential equations in large domains. An iterative numerical algorithm based on a regularising method and the minimization of functionals of the difference between the observations and the numerical solutions of the two-dimensional diffusion equation in the time domain for the determination of the diffusion coefficients is applied. This iterative procedure is derived by using the perturbation method. The DD method of the inverse problem follows the basic ideas of the Schwarz alternating method. Numerical simulations are carried out to test the feasibility of the iterative inversion algorithm with the DD method and to study the intrinsic characteristic of this iterative algorithm with artificially generated data. These numerical results are presented for illustrating the usefulness of the proposed idea. It is found that the iterative method can give good results and the inversion algorithm with DD has the same efficiency as the algorithm without DD.