Observable subspace solution for irreducible POMDPs with infinite horizon.
Lu YuRichard R. BrooksPublished in: CSIIRW (2011)
Keyphrases
- infinite horizon
- partially observable
- markov decision processes
- optimal policy
- optimal control
- dynamic programming
- long run
- finite horizon
- dec pomdps
- partially observable markov decision processes
- stochastic demand
- markov decision process
- reinforcement learning
- production planning
- single item
- state space
- integer programming
- optimal solution
- learning algorithm
- real time
- mathematical model
- markov decision problems
- fixed cost
- average cost
- decision making
- np hard
- decision problems