Error-Bounded Approximations for Infinite-Horizon Discounted Decentralized POMDPs.
Jilles Steeve DibangoyeOlivier BuffetFrançois CharpilletPublished in: ECML/PKDD (1) (2014)
Keyphrases
- infinite horizon
- dec pomdps
- finite horizon
- optimal policy
- optimal control
- long run
- partially observable
- markov decision processes
- dynamic programming
- stochastic demand
- production planning
- lead time
- markov decision process
- average cost
- single item
- state space
- partially observable markov decision processes
- fixed cost
- reinforcement learning
- bayesian networks
- production system
- machine learning