Online Planning for Constrained POMDPs with Continuous Spaces through Dual Ascent.
Arec L. JamgochianAnthony CorsoMykel J. KochenderferPublished in: CoRR (2022)
Keyphrases
- partially observable markov decision processes
- stochastic domains
- belief state
- online learning
- reinforcement learning
- partially observable
- belief space
- dynamic programming
- predictive state representations
- heuristic search
- ai planning
- state space
- decision problems
- continuous state spaces
- continuous domains
- travel planning
- point based value iteration
- stochastic systems
- real time
- decision theoretic
- motion planning
- planning problems
- dynamical systems
- domain independent
- sufficient conditions
- multi agent systems