Constraint-based dynamic programming for decentralized POMDPs with structured interactions.
Akshat KumarShlomo ZilbersteinPublished in: AAMAS (1) (2009)
Keyphrases
- dynamic programming
- dec pomdps
- infinite horizon
- partially observable markov decision processes
- decision theoretic
- reinforcement learning
- continuous state
- multi agent
- theoretical justification
- state space
- single agent
- structured data
- distributed constraint optimization
- optimal plans
- cooperative
- linear programming
- optimal control
- coarse to fine
- planning under uncertainty
- partially observable
- markov decision problems
- real world
- stereo matching
- markov decision processes
- dp matching
- optimal policy
- single machine
- lagrangian relaxation
- dynamic environments
- peer to peer
- np hard
- lower bound
- multiscale