Improved Memory-Bounded Dynamic Programming for Decentralized POMDPs.
Sven SeukenShlomo ZilbersteinPublished in: UAI (2007)
Keyphrases
- dynamic programming
- dec pomdps
- reinforcement learning
- infinite horizon
- partially observable markov decision processes
- multi agent
- distributed constraint optimization
- decision theoretic
- linear programming
- markov decision processes
- continuous state
- single agent
- greedy algorithm
- optimal policy
- theoretical justification
- distributed systems
- improved algorithm
- state space
- bounded memory
- limited memory
- dp matching
- data sets
- single machine
- stereo matching
- peer to peer
- cooperative
- associative memory
- dynamic environments
- policy search
- np hard
- learning algorithm