Recursive structures in the multispecies TASEP

TL;DR

This paper explores recursive structures in the multi-species TASEP, extending matrix product techniques to construct eigenvectors and reveal hierarchical spectral properties across different particle classes.

Contribution

It introduces operators that generalize the matrix ansatz, enabling the construction of eigenvectors for N-TASEP from (N-1)-TASEP eigenvectors, thus revealing new recursive and hierarchical spectral insights.

Findings

Operators for eigenvector construction across classes

Hierarchical spectral inclusion of Markov matrices

Generalization of matrix product representation

Abstract

We consider a multi-species generalization of the totally asymmetric simple exclusion process (TASEP) with the simple hopping rule: for x and yth-class particles (x<y), the transition xy -> yx occurs with a rate independent from the values x and y. P. A. Ferrari and J. Martin (2007) obtained the stationary state of this model thanks to a combinatorial algorithm, which was subsequently interpreted as a matrix product representation by Evans et al. (2009). This `matrix ansatz' shows that the stationary state of the multi-species TASEP with N classes of particles (N-TASEP) can be constructed algebraically by the action of an operator on the (N-1)-TASEP stationary state. Besides, Arita et al. (2009) analyzed the spectral structure of the Markov matrix: they showed that the set of eigenvalues of the N-TASEP contains those of the (N-1)-TASEP and that the various spectral inclusions can be encoded in a hierarchical set-theoretic structure known as the Hasse diagram. Inspired by these works, we define nontrivial operators that allow us to construct eigenvectors of the N-TASEP by lifting the eigenvectors of the (N-1)-TASEP. This goal is achieved by generalizing the matrix product representation and the Ferrari-Martin algorithm. In particular, we show that the matrix ansatz is not only a convenient tool to write the stationary state but in fact intertwines Markov matrices of different values of N.