On transition matrices of Markov chains corresponding to Hamiltonian cycles

Konstantin Avrachenkov, Ali Eshragh, Jerzy Filar

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)

Abstract

In this paper, we present some algebraic properties of a particular class of probability transition matrices, namely, Hamiltonian transition matrices. Each matrix P in this class corresponds to a Hamiltonian cycle in a given graph G on n nodes and to an irreducible, periodic, Markov chain. We show that a number of important matrices traditionally associated with Markov chains, namely, the stationary, fundamental, deviation and the hitting time matrix all have elegant expansions in the first n- 1 powers of P, whose coefficients can be explicitly derived. We also consider the resolvent-like matrices associated with any given Hamiltonian cycle and its reverse cycle and prove an identity about the product of these matrices. As an illustration of these analytical results, we exploit them to develop a new heuristic algorithm to determine a non-Hamiltonicity of a given graph.

Original languageEnglish
Pages (from-to)19-35
Number of pages17
JournalAnnals of Operations Research
Volume243
Issue number1/2
DOIs
Publication statusPublished - 1 Aug 2016

Fingerprint

Dive into the research topics of 'On transition matrices of Markov chains corresponding to Hamiltonian cycles'. Together they form a unique fingerprint.

Cite this