Communication for Improving Policy Computation in Distributed POMDPs.

Ranjit Nair Milind Tambe Maayan Roth Makoto Yokoo

Published in: AAMAS (2004)

Keyphrases

communication overhead
communication cost
partially observable markov decision processes
optimal policy
distributed control
distributed constraint optimization
computer networks
reinforcement learning
distributed systems
cooperative

spatially distributed
fully distributed
multi agent
partially observable
policy gradient
communication networks
belief state
dynamic programming
policy search
distributed computation

markov decision processes
open systems
state space
markov decision problems
multimedia communication
distributed teams
action selection
multi party
point based value iteration
multi agent systems

global knowledge
remote sites
belief space
sensor networks
partial observability
exchange information
peer to peer
markov chain
reward function
decision problems