ORDER-RECURSIVE GAUSSIAN-ELIMINATION (ORGE) AND EFFICIENT CAD OF MICROWAVE CIRCUITS

Planar circuit elements in millimeter-wave integrated circuits (MMIC's ) typically consist of one or more discontinuities (e.g., stubs) conne cted to a number of transmission lines, In the computer-aided design ( CAD) and optimization of such passive elements using the method of mom ents, it is necessary to iteratively simulate many subproblems involvi ng dimensional changes to various parts of the circuit. On examining t he simulation problem closely, it can be seen that there is a consider able overlap of data in various subproblems, In practice, each subprob lem is solved independent of others, without taking into account the d uplication of data, This leads to an inefficient design technique. In this paper, we present a design technique that effectively exploits th e duplication of data by employing a recursive variant of Gaussian eli mination, called order-recursive Gaussian elimination (ORGE). The pote ntial utility of ORGE in microwave circuit simulation and CAD is demon strated by applying it to the design of a microstrip filter.