The effect on linear circuit response of simultaneous changes in a number

of components must often be computed. Two efficient methods for predicting this effect are examined. It is shown that the computational cost of predicting the effect of each set is approximately

, where

is the cost of a Gaussian elimination of an

matrix. However, if the sets represent combinations of relatively few individual component changes, then systematic exploration is shown to reduce the computational cost per set to the order of 2.