𝔖 Bobbio Scriptorium
✦   LIBER   ✦

Ergodic control of partially observed Markov chains

✍ Scribed by Vivek S Borkar

Publisher
Elsevier Science
Year
1998
Tongue
English
Weight
96 KB
Volume
34
Category
Article
ISSN
0167-6911

No coin nor oath required. For personal study only.

✦ Synopsis

The ergodic or long-run average cost control problem for a partially observed ΓΏnite-state Markov chain is studied via the associated fully observed separated control problem for the nonlinear ΓΏlter. Dynamic programming equations for the latter are derived, leading to existence and characterization of optimal stationary policies.

πŸ“œ SIMILAR VOLUMES

Erratum to β€œErgodic control of partially
Erratum to β€œErgodic control of partially observed Markov chains” [Systems Control Lett. 34 (1998) 185–189]
✍ Vivek S Borkar πŸ“‚ Article πŸ“… 1999 πŸ› Elsevier Science 🌐 English βš– 38 KB

at Austin, has pointed out in a private communication a aw in the arguments leading to Lemma 2.3 of [1]. The problem is that (in the notation of [1]) the relation has on the r.h.s. n occurring as an argument of G which introduces a nonlinear dependence of the r.h.s. on n . But then the application

Ergodic decomposition of Markov chains
Ergodic decomposition of Markov chains
✍ CΓ©sar E. Villarreal πŸ“‚ Article πŸ“… 1998 πŸ› Elsevier Science 🌐 English βš– 829 KB

We explicitly find the spectral decomposition, when it exists, of a Markov operator p. :fl ~ El using the asymptotic periodicity of the associated infinite Markov matrix. We give a simple condition under which an infinite Markov matrix is asymptotically periodic. We also determine the set of P'-inva