The expected total cost criterion for Markov decision processes under constraints (Q2856038)
From MaRDI portal
 | This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: The expected total cost criterion for Markov decision processes under constraints |
scientific article; zbMATH DE number 6218390
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | The expected total cost criterion for Markov decision processes under constraints |
scientific article; zbMATH DE number 6218390 |
Statements
23 October 2013
0 references
Markov decision process
0 references
expected total cost criterion
0 references
linear programming
0 references
occupation measure
0 references
0 references
0 references
0 references
0.9761856
0 references
0.9381834
0 references
0.9279606
0 references
0.91904616
0 references
0.91741824
0 references
The expected total cost criterion for Markov decision processes under constraints (English)
0 references
Discrete-time Markov processes (MDPs) with constraints and objectives of the form of expected total cost over the infinite horizon are studied. The problem is analyzed using the linear programming approach. It is shown that if there exists an optimal solution for the associated linear program then there exists a randomized stationary policy which is optimal for the MDP and the optimal value for both problems coincides. Also it is proved that the set of randomized stationary policies provides a sufficient set for solving the MDP. The authors do not assume that the MDP is transient or absorbing and the cost function is nonnegative or bounded below. Three examples that illustrate the obtained results are given.
0 references
