SUFFICIENT CONDITIONS FOR ITERATED TIMING ANALYSIS TO CONVERGE LOCALLY.

Kiichi Urahama

Research output: Contribution to journalArticle

Abstract

This paper investigates the convergence property of a circuit simulation technique called an iterated timing analysis. A block strictly row-wise or column-wise diagonal dominant matrix is defined, and these matrices are shown to be nonsingular. Both block Jacobi method and block Gauss-Seidel iteration are proved to converge for the equation whose coefficient matrix is one of these matrices. On the basis of these results, the following two sufficient conditions for the iterated timing analysis to converge locally are given: (i) the capacitance matrix of the circuit is block strictly row-wise or column-wise diagonally dominant and a time step is sufficiently small, (ii) the conductance matrix of the circuit has the same property and a time step is sufficiently large.

Original languageEnglish
Pages (from-to)1289-1293
Number of pages5
JournalTransactions of the Institute of Electronics and Communication Engineers of Japan. Section E
VolumeE69
Issue number12
Publication statusPublished - Dec 1986

Fingerprint

Networks (circuits)
Circuit simulation
Capacitance

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

SUFFICIENT CONDITIONS FOR ITERATED TIMING ANALYSIS TO CONVERGE LOCALLY. / Urahama, Kiichi.

In: Transactions of the Institute of Electronics and Communication Engineers of Japan. Section E, Vol. E69, No. 12, 12.1986, p. 1289-1293.

Research output: Contribution to journalArticle

@article{5cc0faba641841b2a89bec1aa4ddaec7,
title = "SUFFICIENT CONDITIONS FOR ITERATED TIMING ANALYSIS TO CONVERGE LOCALLY.",
abstract = "This paper investigates the convergence property of a circuit simulation technique called an iterated timing analysis. A block strictly row-wise or column-wise diagonal dominant matrix is defined, and these matrices are shown to be nonsingular. Both block Jacobi method and block Gauss-Seidel iteration are proved to converge for the equation whose coefficient matrix is one of these matrices. On the basis of these results, the following two sufficient conditions for the iterated timing analysis to converge locally are given: (i) the capacitance matrix of the circuit is block strictly row-wise or column-wise diagonally dominant and a time step is sufficiently small, (ii) the conductance matrix of the circuit has the same property and a time step is sufficiently large.",
author = "Kiichi Urahama",
year = "1986",
month = "12",
language = "English",
volume = "E69",
pages = "1289--1293",
journal = "Transactions of the Institute of Electronics and Communication Engineers of Japan. Section E",
issn = "0387-236X",
number = "12",

}

TY - JOUR

T1 - SUFFICIENT CONDITIONS FOR ITERATED TIMING ANALYSIS TO CONVERGE LOCALLY.

AU - Urahama, Kiichi

PY - 1986/12

Y1 - 1986/12

N2 - This paper investigates the convergence property of a circuit simulation technique called an iterated timing analysis. A block strictly row-wise or column-wise diagonal dominant matrix is defined, and these matrices are shown to be nonsingular. Both block Jacobi method and block Gauss-Seidel iteration are proved to converge for the equation whose coefficient matrix is one of these matrices. On the basis of these results, the following two sufficient conditions for the iterated timing analysis to converge locally are given: (i) the capacitance matrix of the circuit is block strictly row-wise or column-wise diagonally dominant and a time step is sufficiently small, (ii) the conductance matrix of the circuit has the same property and a time step is sufficiently large.

AB - This paper investigates the convergence property of a circuit simulation technique called an iterated timing analysis. A block strictly row-wise or column-wise diagonal dominant matrix is defined, and these matrices are shown to be nonsingular. Both block Jacobi method and block Gauss-Seidel iteration are proved to converge for the equation whose coefficient matrix is one of these matrices. On the basis of these results, the following two sufficient conditions for the iterated timing analysis to converge locally are given: (i) the capacitance matrix of the circuit is block strictly row-wise or column-wise diagonally dominant and a time step is sufficiently small, (ii) the conductance matrix of the circuit has the same property and a time step is sufficiently large.

UR - http://www.scopus.com/inward/record.url?scp=0023012621&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0023012621&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0023012621

VL - E69

SP - 1289

EP - 1293

JO - Transactions of the Institute of Electronics and Communication Engineers of Japan. Section E

JF - Transactions of the Institute of Electronics and Communication Engineers of Japan. Section E

SN - 0387-236X

IS - 12

ER -