On number system constructions (Q2460646): Difference between revisions

Let \(\Lambda\) be a lattice, i.e., a free \({\mathbb Z}\)-module of rank \(k\), in \({\mathbb R}^k\), \(M:\Lambda\to\Lambda\) a group endomorphism with \(\det(M)\neq 0\) and \(D\) a finite subset of \(\Lambda\) containing \(0\). The triple \((\Lambda,M,D)\) is called a number system, if every element \(x\in\Lambda\) has a unique, finite representation of the form \(x=\sum_{i=0}^{\ell} M^i d_i\), where \(d_i\in D\) and \(\ell\in{\mathbb N}\). The first half of the paper is devoted to an overview on results in this area. In the second part, the authors prove two results. Let \(\Lambda\) and \(M\) be given. If the spectral radius of \(M^{-1}\) is less than \(1/2\), then there exists a digit set \(D\) for which \((\Lambda, M, D)\) is a number system. Let \(f=c_0+c_1x+\cdots +c_k x^k\in{\mathbb Z}[x]\) with \(c_k=1\) be a polynomial and \(M\) its companion matrix. If the weak dominant condition \(| c_0| >2\sum_{i=1}^k | c_i| \) holds, then there exists a digit set \(D\) for which \(({\mathbb Z}^k, M, D)\) is a number system.