Characterization of cutoff for reversible Markov chains (Q2012242)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Characterization of cutoff for reversible Markov chains |
scientific article |
Statements
Characterization of cutoff for reversible Markov chains (English)
0 references
28 July 2017
0 references
A sequence of Markov chains is said to exhibit (total variation) cutoff if the convergence to stationarity in total variation distance is abrupt. A chain is called lazy if \(P(x,x) \geqslant {1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}\), for all \(x\). Theorem 1 gives a sharp spectral condition for cutoff in lazy weighted nearest-neighbor random walks on trees. Theorem 2 gives conditions for cutoff to exhibit for a generalization of birth and death chains. Theorem 3 considers sequences of lazy reversible irreducible finite chains.
0 references
cutoff
0 references
mixing-time
0 references
finite reversible Markov chains
0 references
hitting times
0 references
trees
0 references
maximal inequality
0 references