Addressing Myopic Constrained POMDP Planning with Recursive Dual Ascent.
Paula StoccoSuhas ChundiArec L. JamgochianMykel J. KochenderferPublished in: ICAPS (2024)
Keyphrases
- planning problems
- belief space
- partially observable markov decision processes
- partially observable markov decision process
- partially observable
- state space
- belief state
- infinite horizon
- reinforcement learning
- partially observable stochastic domains
- partial observability
- planning under uncertainty
- markov decision processes
- dynamical systems
- decision theoretic
- predictive state representations
- stochastic domains
- blocks world
- plan generation
- planning process
- ai planning
- point based value iteration
- finite state
- optimal planning
- least squares
- markov decision problems
- heuristic search
- information gain
- selective perception