Advances and Applications in Statistics
Volume 7, Issue 3, Pages 357 - 388
(December 2007)
|
|
A MAINTENANCE POLICY UNDER MARKOVIAN DETERIORATION
John E. Goulionis (Greece) and V. K. Benos (Greece)
|
Abstract: In this paper, we consider the problem of determining an optimal replacement policy for stochastically deteriorating systems for which only incomplete state information is available. We present a formal definition of the control-limit policy. We provide new sufficient conditions, which ensure the existence of the control-limit policy according to this definition. The structural properties of a three-state partially observable Markov decision process based on these sufficient conditions are derived. Also, we will investigate algorithms for finding the optimal-policy using the total expected discounted cost criterion. The paper closes with a detailed example illustrating the application of the algorithm to the three state partially observable Markov decision process. |
Keywords and phrases: partially observable Markov decision processes, Stochastic Models, Bayesian-Statistics. |
|
Number of Downloads: 360 | Number of Views: 1339 |
|