Topological formulation of symbolic network functions and sensitivity analysis of active networks

A new and efficient algorithm for topological formulation of network functions in symbolic form for a general linear (passive or active) network is presented. This algorithm, which is based on the concept of graph partitioning developed by Dunn and Chan, has all of the advantages of the existing topological formulas, such as overcoming the problem of sign evaluation and greatly reducing the number of term cancellations and the number of topological terms. In addition, and perhaps the most important, it presents a step-by-step procedure for formulating network functions that can readily be adapted for computer implementation. Finally, it is shown that the theoretical results developed herein are most suited for sensitivity analysis of linear active networks.