Polynomial Root-Finding Methods Whose Basins of Attraction Approximate Voronoi Diagram

被引:9
|
作者
Kalantari, Bahman [1 ]
机构
[1] Rutgers State Univ, Dept Comp Sci, Newark, NJ 07102 USA
关键词
Complex polynomials; Voronoi diagram; Zeros; Newton's method; Iteration functions; Fractal; Julia set; Dynamical systems; Computational geometry; FAMILY;
D O I
10.1007/s00454-011-9330-3
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Given a complex polynomial p(z) with at least three distinct roots, we first prove that no rational iteration function exists where the basin of attraction of a root coincides with its Voronoi cell. In spite of this negative result, we prove that the Voronoi diagram of the roots can be well approximated through a high order sequence of iteration functions, the Basic Family, B (m) (z), ma parts per thousand yen2. Let theta be a simple root of p(z), V(theta) its Voronoi cell, and A (m) (theta) its basin of attraction with respect to B (m) (z). We prove that given any closed subset C of V(theta), including any homothetic copy of V(theta), there exists m (0) such that for all ma parts per thousand yenm (0), C is also a subset of A (m) (theta). This implies that when all roots of p(z) are simple, the basins of attraction of B (m) (z) uniformly approximate the Voronoi diagram of the roots to within any prescribed tolerance. Equivalently, the Julia set of B (m) (z), and hence the chaotic behavior of its iterations, will uniformly lie to within prescribed strip neighborhood of the boundary of the Voronoi diagram. In a sense, this is the strongest property a rational iteration function can exhibit for polynomials. Next, we use the results to define and prove an infinite layering within each Voronoi cell of a given set of points, whether known implicitly as roots of a polynomial equation, or explicitly via their coordinates. We discuss potential application of our layering in computational geometry.
引用
收藏
页码:187 / 203
页数:17
相关论文
共 50 条
  • [31] On the order of convergence of a determinantal family of root-finding methods
    Kalantari, B
    BIT NUMERICAL MATHEMATICS, 1999, 39 (01) : 96 - 109
  • [32] On a family of parallel root-finding methods for generalized polynomials
    Huang, ZD
    Zheng, SM
    APPLIED MATHEMATICS AND COMPUTATION, 1998, 91 (2-3) : 221 - 231
  • [33] Survey of several root-finding methods in the complex plane
    Szyszkowicz, Mieczyslaw
    Computer Graphics Forum, 1991, 10 (02) : 141 - 144
  • [34] Quasi-analytical root-finding for non-polynomial functions
    Cleonice F. Bracciali
    Michael Carley
    Numerical Algorithms, 2017, 76 : 639 - 653
  • [35] Backward stability of polynomial root-finding using Fiedler companion matrices
    De Teran, Fernando
    Dopico, Froilan M.
    Perez, Javier
    IMA JOURNAL OF NUMERICAL ANALYSIS, 2016, 36 (01) : 133 - 173
  • [36] On Some Optimal Multiple Root-Finding Methods and their Dynamics
    Kansal, Munish
    Kanwar, V.
    Bhatia, Saurabh
    APPLICATIONS AND APPLIED MATHEMATICS-AN INTERNATIONAL JOURNAL, 2015, 10 (01): : 349 - 367
  • [37] On the Order of Convergence of a Determinantal Family of Root-Finding Methods
    Bahman Kalantari
    BIT Numerical Mathematics, 1999, 39 : 96 - 109
  • [38] Computational Geometry as a Tool for Studying Root-Finding Methods
    Petkovic, Ivan
    Rancic, Lidija Z.
    FILOMAT, 2019, 33 (04) : 1019 - 1027
  • [39] Interval root-finding methods of Laguerre's type
    Petkovic, LD
    Petkovic, MS
    Zivkovic, D
    INCLUSION METHODS FOR NONLINEAR PROBLEMS: WITH APPLICATIONS IN ENGINEERING, ECONOMICS AND PHYSICS, 2003, 16 : 199 - 211
  • [40] Quasi-analytical root-finding for non-polynomial functions
    Bracciali, Cleonice F.
    Carley, Michael
    NUMERICAL ALGORITHMS, 2017, 76 (03) : 639 - 653