Generalized and bounded policy iteration for finitely-nested interactive POMDPs: scaling up.
Ekhlas SonuPrashant DoshiPublished in: AAMAS (2012)
Keyphrases
- policy iteration
- markov decision processes
- reinforcement learning
- optimal policy
- policy iteration algorithm
- markov decision problems
- average reward
- sample path
- model free
- partially observable
- markov decision process
- finite state
- state space
- policy evaluation
- partially observable markov decision processes
- temporal difference
- dynamic programming
- fixed point
- infinite horizon
- optimal control
- least squares
- multi agent
- learning algorithm
- machine learning
- linear programming
- reinforcement learning algorithms
- average cost
- finite number
- markov chain
- continuous state
- np hard
- actor critic