Theoretical results on the effect of 'shortcut' actions in MDPs.
Sara M. McCarthyDoina PrecupPublished in: Connect. Sci. (2014)
Keyphrases
- markov decision processes
- decision theoretic planning
- reinforcement learning
- state and action spaces
- action space
- initial state
- partially observable
- situation calculus
- reward function
- optimal policy
- stochastic domains
- dynamic programming
- state transitions
- decision theoretic
- theoretical analysis
- action sets
- plan recognition
- decision making
- average cost
- decision problems
- reasoning about actions
- multiagent reinforcement learning
- factored mdps