Contragredient transformations applied to the optimal projection equations (Q1260811)

It is shown how the optimal projection equations for solving the \(H_ 2\) model reduction problem can be transformed, by contragredient transformation, into forms suitable for algorithms for solving nonlinear problems. An algorithm based on the singular value decomposition for the contragredient transformation is derived. Three possible ways of transforming the optimal projection equations, using contragredient transformations, into a computationally useful form are given. A numerical homotopy algorithm for the transformed equations is described and a numerical example is given.