Abstracting Complex Domains Using Modular Object-Oriented Markov Decision Processes.
Shawn SquireMarie desJardinsPublished in: AAAI (2016)
Keyphrases
- markov decision processes
- complex domains
- object oriented
- reinforcement learning
- optimal policy
- state space
- finite state
- dynamic programming
- transition matrices
- relational databases
- domain theory
- planning under uncertainty
- bayesian networks
- policy iteration
- decision theoretic planning
- reachability analysis
- function approximation
- partially observable
- average reward
- factored mdps
- model free
- action space
- artificial intelligence
- finite horizon
- markov decision process
- reward function
- infinite horizon
- state and action spaces
- optimal control
- machine learning
- model based reinforcement learning