Bellman inequalities in markov decision deterministic drift processes (Q3786304)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 4049426
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Bellman inequalities in markov decision deterministic drift processes |
scientific article; zbMATH DE number 4049426 |
Statements
Bellman inequalities in markov decision deterministic drift processes (English)
0 references
1987
0 references
Bellman inequalities
0 references
quasi-variational inequalities
0 references
Bellman optimality equation
0 references
Markov decision deterministic drift processes
0 references
\(\epsilon \)- optimal policies
0 references
randomly split time parameter
0 references
0 references
0.8107112050056458
0 references
0.7627613544464111
0 references
0.7572956681251526
0 references
0.754104733467102
0 references
In the case of both continuously and impulsively controlled Markov-type stochastic processes the usual Bellman optimality equation for the value function v turns into a system of the form \(T^ 1v\leq 0\), \(T^ 2v\leq 0,T^ 2v\cdot T^ 2=0\), where \(T^ 1\) and \(T^ 2\) are, respectively, a non-differential and a differential nonlinear operator (in the controlled diffusion this system became known as ``quasi-variational inequalities''). We examine to what extent analogous inequalities are valid in Markov decision deterministic drift processes, in which randomness affects only jumps. As auxiliary results we prove that v is upper semianalytic, that almost everywhere \(\epsilon\)-optimal policies do exist, and also prove a defect formula for policies which is closely related to the inequalities in consideration. Our investigation is based on the concept of stochastic processes with a randomly split time parameter.
0 references