Spectral properties of zero temperature dynamics in a model of a compacting granular column (Q2429392)

Zero temperature dynamics is dominated by constraints and lends itself to non-generic behavior. The authors focus on a model of granular compaction in order to see that the peculiarities of the allowed motions induce non-exponential relaxation even when all eigenvalues for the stochastic dynamics are real. This characteristic behavior has a mathematical mechanism that lies in the fact that the associated eigenvectors do not span the entire state space and the best one can do is to represent the generator of the stochastic dynamics as a Jordan form. Granular material can exhibit features significantly different from those of traditional fluids. We have models describing such a situation (see References 9-13) but the proposed models use observable representation that can be used for non-zero temperature dynamics. However, there are several problems of applications when the constraints of zero temperature are imposed: the stochastic matrix generated under these rules does not have a full complement of eigenvectors. In this paper a development of the stochastic matrix governing this process is presented. The text is divided into ten sections. After the preliminaries, in Section 2 we have a presentation of the matrix generator of the stochastic dynamics that implements the rules given in the introduction. In Section 3 the authors examine the eigenvectors. Then they consider the anomalous time-dependence. In Section 5 a proposed model of dynamics is given from a different perspective: the independence of spin-\(k\) dynamics. All theoretical considerations given so far are illustrated in Section 6 by non-trivial examples. After this we have an interesting discovery -- the zero-temperature dynamics does not always lead to the Jordan form. This is explained in the next section, and another perspective on this problem is given in Section 8. In the next section the authors show that the basis for diagonalizing the hypercube random walk brings the stochastic dynamics matrix to a triangular form, thereby unifying the two approaches, that of correlations and that of algebraic recursions. The last section discusses the presented results.