Generation of equivalent active networks with minimized sensitivities

A method for generating equivalent networks with minimum sensitivities is presented, applicable to both passive and active linear time invariant networks. Special transformation matrices can modify the element values and topology, and can grow elements into the network. First-order sensitivities of the transformed networks are obtained and subjected to minimization. As a result, networks with a larger number of components but smaller individual sensitivities can be obtained. The method is demonstrated on active networks.