Abstract
In this thesis a study of a Semi-Direct method for the solution of large scale circuits is presented. The method combines features from the Newton- Raphson method and the Gauss-Seidel method. These two methods are both illustrated. The Semi-Direct method is presented both theoretically and empirically using three programs developed for this purpose. The Semi-Direct method and the Newton-Raphson method are compared. The comparison includes speed (number of iterations and execution time) and storage requirements. The Semi-Direct method definitely has storage advantages over the Newton-Raphson method at all circuit levels of 2 or more nodes. If some set conditions are met in the circuit, the Semi-Direct method will require less CPU time to reach the solution than the Newton-Raphson method
Date of this Version
12-1-1984
Comments
MS thesis. This work was supported by: International Business Machines Corporation