Harbrecht, Helmut; Tausch, Johannes: An efficient numerical method for a shape-identification problem arising from the heat equation, , 2011

This paper is dedicated to the determination of the shape of a compactly supported constant source in the heat equation from measurements of the heat flux through the boundary. This shape-identification problem is formulated as the minimization of a least-squares cost functional for the desired heat flux at the boundary. The shape gradient of the shape functional under consideration is computed by means of the adjoint method. A gradient-based nonlinear Ritz–Galerkin scheme is applied to discretize the shape optimization problem. The state equation and its adjoint are computed by a fast space-time multipole method for the heat equation. Numerical experiments are carried out to demonstrate the feasibility and scope of the present approach.

Harbrecht, Helmut; Schwab, Christoph: Sparse tensor finite elements for elliptic multiple scale problems, , 2011

 

Locally periodic, elliptic multiscale problems in a bounded Lipschitz domain $Dsubsetmathbb{R}^n$ with $Kgeq 2$ separated scales are reduced to an elliptic system of $K$ coupled, anisotropic elliptic one-scale problems in a cartesian product domain of total dimension $Kn$ (e.g.~cite{N89,A,AB,CDG2008}). In cite{HS,SSt08}, it has been shown how thesecoupled elliptic problems could be solved by sparse tensor wavelet Finite Element methods in log-linear complexity with respect to the number $N$ of degrees of freedom required by multilevel solvers for elliptic one-scale problems in $D$ with the same convergence rate.In the present paper, the high dimensional one-scale limiting problems are discretized by a sparse tensor product finite element method (FEM) with standard, one-scale FE basis functions as used in engineering FE codes. Sparse tensorizationand multilevel preconditioning is achieved by a BPX multilevel iteration.
We show that the resulting sparse tensor multilevel FEM resolves all physical length scales throughout the domain, with efficiency (i.e.,~accuracy versus work and memory) comparable to that of multigrid solvers for elliptic one-scale problems in the physical domain $D$. In particular, our sparse tensor FEM gives numerical approximations of the correct homogenized limit as well as compressed numerical representationsof all first order correctors, throughout the physical domain with performance independent of the physical problem's scale parameters. Numerical examples with standard FE shape functions and BPX multilevelpreconditioners for elliptic problems with $K=2$ separated, physical scales in spatial dimension $n=2$ confirm the theoretical results. In particular, the present approach allows to avoid the construction of wavelet FE bases necessaryin previous work cite{HS,SSt08} while achieving resolution of all scales throughoutthe physical domain in log-linear complexity, with the logarithmic exponent behaving linearly in the number $K$ of scales.