scientific article; zbMATH DE number 3891132
bang-bang optimal policiesexistence of an optimal randomized stationary policyfinite state and compact action spacelong-run average reward
Existence of optimal solutions belonging to restricted classes (Lipschitz controls, bang-bang controls, etc.) (49J30) Dynamic programming in optimal control and differential games (49L20) Numerical methods involving duality (49M29) Dynamic programming (90C39) Markov and semi-Markov decision processes (90C40) Optimal stochastic control (93E20)
[For the entire collection see Zbl 0538.00036.] The author investigated Markov decision problems with long-run average reward, when there is a global constraint on the long-run average. Assuming finite state and compact action space, the existence of an optimal randomized stationary policy was proved. Moreover, using the Lagrangian technique, he derived dynamic programming equations and obtained conditions for the existence of almost bang-bang optimal policies.
- scientific article; zbMATH DE number 1348599
- On constrained Markov decision processes
- Markov decision programming with constraints
- Constrained Markovian decision processes: The dynamic programming approach
- Markov decision processes under observability constraints
- Constrained Markov decision processes with expected total reward criteria
- scientific article; zbMATH DE number 176107
- Constrained Markov decision processes with uncertain costs
- Non-homogeneous Markov decision processes with a constraint
- Markov decision processes with a constraint on the asymptotic failure rate
- A policy improvement method for constrained average Markov decision processes
- On constrained Markov decision processes
- Deterministic optimal policies for Markov control processes with pathwise constraints
- Markov Decision Processes with Asymptotic Average Failure Rate Constraint
- The expected total cost criterion for Markov decision processes under constraints: a convex analytic approach
- Dynamic programming in constrained Markov decision processes
- Constrained Undiscounted Stochastic Dynamic Programming
- scientific article; zbMATH DE number 4053502 (Why is no real title available?)
- Randomized and Past-Dependent Policies for Markov Decision Processes with Multiple Constraints
- scientific article; zbMATH DE number 176107 (Why is no real title available?)
- scientific article; zbMATH DE number 465368 (Why is no real title available?)
- On Bellman's principle with inequality constraints
- Constrained markov decision processes with compact state and action spaces: the average case
- Non-ergodic Markov decision processes with a constraint on the asymptotic failure rate: general class of policies
- Markov Decision Processes with Sample Path Constraints: The Communicating Case
- Concurrent MDPs with Finite Markovian Policies
- Constrained Markov decision processes with expected total reward criteria
- The Markov Transition Constraint
- On Markov policies for minimax decision processes
- Optimal policies for constrained average-cost Markov decision processes
- Markov control processes with pathwise constraints
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3347664)