Optimality of the NVI adaptive policy for a partially observed Markov decision model

Optimality of the NVI adaptive policy for a partially observed Markov decision model

0.00 Avg rating0 Votes
Article ID: iaor19931709
Country: Poland
Volume: 20
Issue: 4
Start Page Number: 79
End Page Number: 98
Publication Date: Jan 1991
Journal: Control and Cybernetics
Authors: ,
Keywords: production, markov processes
Abstract:

The production replacement process is modelled with partially observed Markov process. The non-stationary value iteration (NVI) adaptive policy for the case of average cost as performance index is derived and its optimality is demonstrated. A parameter estimation algorithm, developed by the same authors, and used for specification of the adaptive policy, is also described. Numerous previously assumed conditions are relaxed or eliminated.

Reviews

Required fields are marked *. Your email address will not be published.