Large change sensitivity analysis in linear systems using generalized householder formulae
J.W. Bandler, Q. J. Zhang · International Journal of Circuit Theory and Applications · 1986
Abstract This paper investigates multiparameter large change sensitivity problems in linear systems by a set of generalized Householder formulae. The newly developed rectangular formulae can accommodate large, small and zero parameter changes directly by avoiding a critical matrix inversion as compared to the traditional square formulae. Possible determination of a minimum order reduced system, whose solution procedure constitutes the major work in large change evaluation is discussed. Applications to linear systems are considered for the original and adjoint systems w.r.t. single as well as multiple input‐output cases. This approach makes it possible to use large change analysis algorithms even if many parameters are changed.