RTDP: A Belief Branch and Bound Real-Time Dynamic Programming Approach to Solving POMDPs.
Sigurdur O. AdalgeirssonCynthia BreazealPublished in: CoRR (2022)
Keyphrases
- real time dynamic programming
- branch and bound
- markov decision processes
- state space
- lower bound
- search algorithm
- branch and bound algorithm
- search space
- optimal solution
- upper bound
- combinatorial optimization
- belief revision
- belief functions
- belief state
- beam search
- markov decision problems
- tree search
- search strategies
- dynamic programming
- expected utility
- heuristic search
- finite state
- reinforcement learning
- learning algorithm
- metaheuristic
- optimal policy
- supervised learning