Given a parameterized model, the RBM determines self adaptively where exact, so-called

snapshot solutions of the full problem are taken. These snapshot solutions form the reduced

basis and are used to construct a reduced model in the offline phase. The reduced model is

obtained by Galerkin projection, which could be characterized as physical interpolation: the

solution for a given point in the parameter space is not obtained by data interpolation, but

by solution of an electromagnetic scattering problem on the reduced basis space. The RBM is applicable in challenging real-time measurement environments as the semiconductor chip production.