On large deviation theorems for Markov processes on a compact domain

大倉, 弘之, OKURA, Hiroyuki

type:Article

Large deviation theorems of the Donsker-Varadhan type are studied. Those theorems for a family of stochastic processes converging to a Markov process have already been obtained by the present author. In this paper, these theorems are modified so that they cover the case where each process is killed on exiting a compact domain. The general theory of large deviations of such a type is mentioned at two levels; the state-space level and the path-space level, and applied to the study of the principal eigenvalue of the generator of a Markov process. It is shown that the principal eigenvalue converges as the probability law of the corresponding Markov process converges. As a typical example, the converging family of Markov processes in the homogenization problem is investigated.

identifier:京都工芸繊維大学　工芸学部研究報告　第４８巻　理工・欧文(1999) pp.11-22

On large deviation theorems for Markov processes on a compact domain

説明

キーワード

詳細情報詳細情報について

書き出し

問題の指摘

On large deviation theorems for Markov processes on a compact domain

説明

キーワード

詳細情報 詳細情報について

書き出し

問題の指摘

詳細情報詳細情報について