A modified Newton method for polynomials
暂无分享,去创建一个
A modified Newton method for polynomials is discussed. It is assumed one has approximations for all the roots of the polynomial. Three variations are described. If the roots are simple, it is shown that under appropriate conditions, two of the variations are cubically convergent, 1. ! ,l t r o d u e t io, l,et t:.here be given film polynomial J'(:C) = aMt(' @ a,:.-.tX"' @ "'" + aix-I-ao (1) ,i i,! where we assume the roots are distinct. Assume X~0), ... , ,~,, tare n distinct guesses for the roots of f(x).
[1] W. Börsch-Supan. A posteriori error bounds for the zeros of polynomials , 1963 .
[2] Immo O. Kerner,et al. Ein Gesamtschrittverfahren zur Berechnung der Nullstellen von Polynomen , 1966 .
[3] Hans J. Maehly,et al. Zur iterativen Auflösung algebraischer Gleichungen , 1954 .
[4] Xiaomei Yang. Rounding Errors in Algebraic Processes , 1964, Nature.
[5] Immo O. Kerner,et al. Algorithm 283: Simultaneous displacement of polynomial roots if real and simple , 1966, Commun. ACM.