On the existence of stationary optimal policies in discrete dynamic programming.
Karel SladkýPublished in: Kybernetika (1981)
Keyphrases
- optimal policy
- dynamic programming
- stationary policies
- markov decision processes
- state space
- infinite horizon
- finite horizon
- markov decision process
- multistage
- reinforcement learning
- finite state
- dynamic programming algorithms
- decision problems
- average cost
- non stationary
- long run
- state dependent
- linear program
- markov decision problems
- linear programming
- serial inventory systems
- bayesian reinforcement learning
- initial state
- total reward
- optimal control
- control policies
- average reward
- lost sales
- policy iteration
- machine learning
- sufficient conditions
- multi agent
- partially observable markov decision processes
- lagrangian relaxation