Families of periodic Jacobi-Perron algorithms for all period lengths (Q301462)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Families of periodic Jacobi-Perron algorithms for all period lengths |
scientific article |
Statements
Families of periodic Jacobi-Perron algorithms for all period lengths (English)
0 references
30 June 2016
0 references
The author studies the Jacobi-Peron algorithm (JPA), for the simultaneous approximation of \(n\) real numbers (the well known generalization of the ordinary continued fraction expansion for a real number [\textit{C. G. J. Jacobi}, J. Reine Angew. Math. 69, 29--64 (1868; JFM 01.0062.01)] \([n=2]\) and [\textit{O. Perron}, Math. Ann. 64, 1--76 (1907; JFM 38.0262.01)] [arbitrary \(n\geq 2\)]). The focus is periodicity and the Hasse-Bernstein unit, referring to results by \textit{C. Levesque} and \textit{G. Rhin} [J. Number Theory 37, No. 2, 173--180 (1991; Zbl 0723.11032)] and \textit{B. Adam} [Math. Comput. 64, No. 212, 1687--1704 (1995; Zbl 0858.11070)]. The main result is Theorem 1. For all integers \(c_0,c_1,\ldots,c_{n-1}m,n,t\) with \(c_{n-1}\geq c_0=1,c_{n-1}\geq c_{n-2}\geq \cdots \geq c_1\geq 0,m\geq n\geq 2\), and \(t\geq 1\), let \[ f(X)=X^n-(u_{n-1,m}t+c_{n-1})X^{n-1}-\cdots -(u_{0,m}t+c_0), \] and let \(\omega\) be the unique positive root of \(f(X)\). The JPA expansion of \(\vec{\alpha}^{(0)}\) with \[ \alpha_i^{(0)}=\sum_{j=0}^i\,(u_{j,m}t+c_j)\omega^{j-i},\;i=1,2,\ldots,n-1, \] is purely periodic of period length \(m\). {ıthe author then gives the partial numerators and denominators of the continued fraction explicitly} Furthermore, the Hasse-Bernstein unit associated with the JPA is \[ \varepsilon=u_{n-1,m}\omega^{n-1}+u_{n-2,m}\omega^{n--2}+\cdots +a_{1,m}\omega+u_{0,m}. \] The numbers used in the algorithm are introduced starting with any \(n\)-tuple \((c_0,c_1,\ldots,c_{n-1})\) of non-negative integers with \(c_0,c_{n-1}\geq 1\). These are used to introduce \(n\) related recurrence sequences by \[ u_{i,m}=c_{n-1}u_{i,m-1}+c_{n-2}u_{i,m-2}+\cdots +c_0u_{i,m-n},\;0\leq i\leq n-1,\;m\geq n. \] The proof follows after a chain of lemmas.
0 references
Jacobi-Perron algorithms
0 references
multi-dimensional continued fractions
0 references
periodic expansions
0 references
0 references