Mat-1.3656 Seminar on
numerical analysis and computational science
Monday Nov 10, 2008, room
U356 at 14.15, Eirola & Stenberg
Toni Lassila, TKK Mat & EPFL
Convex shape identification from a single interferogram
An inverse problem in low temperature physics asks to reconstruct a crys-
tal shape based on a single interferometer measurement. It is shown that
the inverse problem is uniquely solvable if the crystal shape is required to
be convex and its boundary is assumed to be known.
Numerical 1-d examples based on the gradient descent method for shape
optimization are presented. Some of the difficulties encountered when per-
forming variational optimization in the class of convex functions are dis-
cussed.