Compact, Convex Upper Bound Iteration for Approximate POMDP Planning.
Tao WangPascal PoupartMichael H. BowlingDale SchuurmansPublished in: AAAI (2006)
Keyphrases
- upper bound
- point based value iteration
- planning problems
- lower bound
- partially observable markov decision processes
- belief space
- partially observable
- piecewise linear
- partially observable markov decision process
- belief state
- stochastic domains
- worst case
- state space
- reinforcement learning
- provide an upper bound
- planning under uncertainty
- branch and bound algorithm
- policy evaluation
- newton raphson
- decision theoretic
- convex hull
- lower and upper bounds
- predictive state representations
- optimal policy
- partially observable stochastic domains
- partial observability
- markov decision problems
- approximate solutions
- dynamic programming
- convex optimization
- dynamical systems
- sequential decision making problems
- heuristic search
- ai planning
- multi agent
- partially observable stochastic games
- infinite horizon
- domain independent
- np hard
- optimal solution
- objective function