Expansion of vectors by powers of a matrix (Q1036963)

The authors generalize the standard expansion of an integer \(z=d_{N+1}n^{N+1}+\sum_{j=0}^N d_jn^j\), \(d_{N+1}\in \{0,-1\}\), \(d_j\in \{0,\dots,n-1\}\) to the vector case so that \(n\) is replaced by a \(d\times d\)-matrix \(M\) with integer elements and eigenvalues greater than \(1\) in absolute value and show that every \(x\in {\mathbb Z}^d\) can be written in the form \[ x=M^{N+1}z_0 + \sum_{k=0}^N M^ky_k, \] where the \(y_k\) belong to a set, with \(\mathrm {det}(M)\) elements, of so-called \(M\)-digits and \(z_0\) is either zero or such that \(z_0=M^rz_0+M^{r-1}s_{r-1}+M^{r-2}s_{r-2}+\dots+Ms_1+s_0\) where \(s_0,s_1,\dots ,s_{r-1}\) are \(M\)-digits. Thus integer vectors are divided into \(M\)-positive and \(M\)-negative ones depending on whether \(z_0\) is zero or not and several examples of how \({\mathbb Z}^2\) is divided into \(M\)-positive and \(M\)-negative vectors are given. A similar expansion of real vectors is given as well.