Patterns generation and transition matrices in multi-dimensional lattice models (Q2569354)

This article deals with the problem how to count multi-color patterns on a large rectangular array which are generated from a set of smaller, admissible patterns. The approach of the authors is based on a careful enumeration of all patterns. This allows one in the case of a two-dimensional lattice with admissible \(2\times 2\)-patterns to define horizontal and vertical transition matrices. Their relation to Lindenmayer systems and Hadamard matrix products is shortly mentioned. The connection to lattice dynamical systems is made via the spatial entropy which for two-dimensional rectangular \((N_1\times N_2)\)-lattices is defined as \[ h(\mathcal{B})=\lim_{N_1\cdot N_2\to\infty}\frac{\log \Sigma(\mathcal{B})}{N_1\cdot N_2}, \] where \(\Sigma(\mathcal{B})\) is the number of different admissible patterns generated from a basic set \(\mathcal{B}\) of \(2\times 2\)-patterns. As in the one-dimensional case, the spatial entropy is then closely related to the largest eigenvalue of the transition matrices. Since the size of the transition matrices grows exponentially, it is often not easy to determine the asymptotic behaviour of the largest eigenvalue. However, for some special cases, the authors succeed to determine the eigenvalues and the associated spatial entropy.