The algorithm for L-2-optimal model reduction described in Krajewski et al.
(1995), though computationally efficient, may sometimes fail to converge.
In this note, it is shown that, by exploiting the bounds on the eigenvalues
of the Jacobian of the associated transition function, a variant can be de
veloped whose convergence to the minima of the considered index is guarante
ed. The resulting algorithm requires a small additional amount of computati
on with respect to the original procedure. (C) 1999 Elsevier Science Ltd. A
ll rights reserved.