The permutable POMDP: fast solutions to POMDPs for preference elicitation.
Finale DoshiNicholas RoyPublished in: AAMAS (1) (2008)
Keyphrases
- preference elicitation
- partially observable markov decision processes
- reinforcement learning
- belief state
- point based value iteration
- continuous state
- utility function
- partially observable
- belief space
- finite state
- optimal policy
- markov decision processes
- decision problems
- dec pomdps
- dynamical systems
- state space
- approximate solutions
- partially observable markov decision process
- multi criteria
- dynamic programming
- inverse reinforcement learning
- multi dimensional
- minimax regret
- decision theory
- machine learning
- desirable properties
- search space
- multi agent
- objective function
- decision making