Discretized Approximations for POMDP with Average Cost
Huizhen YuDimitri P. BertsekasPublished in: CoRR (2012)
Keyphrases
- average cost
- finite state
- markov decision processes
- optimal policy
- policy evaluation
- infinite horizon
- partially observable markov decision processes
- markov decision process
- markov decision problems
- markov chain
- partially observable
- policy iteration
- finite horizon
- long run
- reinforcement learning
- approximate dynamic programming
- state space
- markov decision chains
- finite number
- optimal control
- decision problems
- risk sensitive
- model checking
- multistage
- dynamic programming
- control policy
- inventory models
- initial state
- action sets
- steady state
- average reward
- control policies
- least squares
- total cost
- machine learning
- lower bound
- stationary policies
- approximation methods
- state dependent
- linear program
- belief state