A State-Space Acyclicity Property for Exponentially Tighter Plan Length Bounds.
Mohammad AbdulazizCharles GrettonMichael NorrishPublished in: ICAPS (2017)
Keyphrases
- state space
- upper bound
- lower bound
- markovian decision
- heuristic search
- goal state
- upper and lower bounds
- planning tasks
- reinforcement learning
- markov decision processes
- monte carlo
- initial state
- lipschitz continuity
- optimal policy
- markov chain
- planning graph
- dynamical systems
- state variables
- dynamic programming
- worst case
- classical planning
- reinforcement learning algorithms
- branch and bound
- plan recognition
- orders of magnitude
- search space
- heuristic function
- particle filter
- decision theoretic
- belief state
- maximum number
- partially observable
- machine learning
- np hard
- query processing
- partial plans
- continuous time markov process
- database systems