Large Scale Optimization of Analog Circuits with Microwave Applications

. This thesis addresses itself to computer oriented techniques f<?r large scale optimization of analog circuits. ~ew techniques for simulutil1n and sensitivity analysis are described and are .used .to improve t?e performance of circuit optimizati~n. A powerful autoljlatic dec'omposition technique \s developed directly 'enabling a normal optim~zer to solve large circuit problems. Our theory is applied to , the design of microwave circuits. "The statU5 of large scale circuit optimization and the state-of-the·art of , microwave CAD arc reviewed. The necessity of circuit oriented .optimization techniques is demonstrated ~'formulatingdesign, modelling, diaL;Tlosis and.tuning into , . ~ optimization problems. A comprehensive treatment of large chan~e sensitivity computation for linearized circuits using generalized Householder formuius is presentcd. :\ technique for circuit response updating via a mini'mum ~rder reduced sy~;tel'n is devciopcd. By avoiding re-analysis of the complete circuit, our method is responsiblc for efficient simulation of large circuits when a subset of the circuit parameters is frequently perturbed. An elegant theory for simulation and exact scnsitivity analysis of branched cascaded nct\\:orks is described. Our approach explicitly takes the circuit-structure -iNo consideration and does not detcriorate as the overall network becl)~cs lan~e The .practicality of the theory is illustrated by efficient optimization of microwavc multiplexers consisting of multi'-cavit3( filters distributed alon~ a wave~uidc manif.)ld. Examples ofoptimizing 12and 16-channcl multiplexers arc provided