Title :
New relaxation-based circuit simulator based on fast automatic differentiation
Author :
Kumar, Girish ; Yamauchi, Masayuki ; Otake, Tsuyoshi ; Tanaka, Mamoru
Author_Institution :
Dept. of Electr. & Electron. Eng., Sophia Univ., Tokyo, Japan
Abstract :
This paper describes the Sophia relaxation-based program (SRP) which is a new circuit simulator with fast automatic differentiation (FAD) on Visual C++ studio. The SRP adds the FAD object of the differentiation for a function f(x) as its differential elements in a Jacobian matrix. This means that the user defines only the description of many nonlinear functions in the input file. Also, in the SRP, the new relaxation method is used to analyze the linearized equation for Newton Raphson method (NRM). Especially the SRP will be used efficiently to analyze cellular neural networks with many templates and many different nonlinear functions. The simulation and the experiment results are shown for analysis of a cellular oscillating nonlinear network.
Keywords :
C++ language; Jacobian matrices; circuit analysis computing; differentiation; nonlinear network analysis; oscillators; relaxation theory; Jacobian matrix; Newton Raphson method; Sophia relaxation-based program; Visual C++ studio; cellular neural network analysis; cellular oscillating nonlinear network analysis; differential element; fast automatic differentiation; linearized equation analysis; nonlinear function definition; relaxation-based circuit simulator; Analytical models; Cellular networks; Cellular neural networks; Circuit simulation; Jacobian matrices; Large-scale systems; Nonlinear equations; Relaxation methods; SPICE; Very large scale integration;
Conference_Titel :
Circuits and Systems, 2004. ISCAS '04. Proceedings of the 2004 International Symposium on
Print_ISBN :
0-7803-8251-X
DOI :
10.1109/ISCAS.2004.1328295
