RUS  ENG
Full version
JOURNALS // Izvestiya Rossiiskoi Akademii Nauk. Seriya Matematicheskaya // Archive

Izv. RAN. Ser. Mat., 2015 Volume 79, Issue 2, Pages 101–136 (Mi im8198)

This article is cited in 2 papers

The spectral method and ergodic theorems for general Markov chains

S. V. Nagaev

Sobolev Institute of Mathematics, Siberian Branch of the Russian Academy of Sciences, Novosibirsk

Abstract: We study the ergodic properties of Markov chains with an arbitrary state space and prove a geometric ergodic theorem. The method of the proof is new: it may be described as an operator method. Our main result is an ergodic theorem for Harris–Markov chains in the case when the return time to some fixed set has finite expectation. Our conditions for the transition function are more general than those used by Athreya–Ney and Nummelin. Unlike them, we impose restrictions not on the original transition function but on the transition function of an embedded Markov chain constructed from the return times to the fixed set mentioned above. The proof uses the spectral theory of linear operators on a Banach space.

Keywords: embedded Markov chain, uniform ergodicity, resolvent, spectral method, stationary distribution.

UDC: 519.21+517.98

MSC: 60J10, 47A35

Received: 16.12.2013
Revised: 20.10.2014

DOI: 10.4213/im8198


 English version:
Izvestiya: Mathematics, 2015, 79:2, 311–345

Bibliographic databases:


© Steklov Math. Inst. of RAS, 2025