Path decompositions for Markov chains. (Q1879829)

The decomposition of a Brownian motion [see \textit{D. Williams}, Proc. Lond. Math. Soc., III. Ser. 28, 738--768 (1974; Zbl 0326.60093); \textit{J. W. Pitman}, Adv. Appl. Probab. 7, 511--526 (1975; Zbl 0332.60055)] has been the starting point for further investigations [see e.g. \textit{J. Bertoin}, Stochastic Processes Appl. 47, 17--35 (1993; Zbl 0786.60101); \textit{P. Greenwood} and \textit{J. W. Pitman}, Adv. Appl. Probab. 12, 893--902 (1980; Zbl 0443.60037); \textit{P. W. Millar}, Trans. Am. Math. Soc. 226, 365--391 (1977; Zbl 0381.60062); \textit{H. Tanaka}, Tokyo J. Math. 13, 375--389 (1990; Zbl 0734.60075)]. In this paper the authors introduce a general method of path decomposition Markov chains by means of positive harmonic functions (Theorems 2 and 3), which covers some of the mentioned results. They affirm that the case of continuous time will be treated elsewhere. Some examples involved random walks and the Pólya urn are given.