Knowledge-Based Policies for Qualitative Decentralized POMDPs.
Abdallah SaffidineFrançois SchwarzentruberBruno ZanuttiniPublished in: AAAI (2018)
Keyphrases
- partially observable markov decision processes
- dec pomdps
- distributed constraint optimization
- optimal policy
- reinforcement learning
- multi agent
- policy search
- finite state
- policy gradient methods
- dynamic programming
- dynamical systems
- continuous state
- markov decision problems
- decision making under uncertainty
- predictive state representations
- markov decision processes
- quantitative and qualitative
- belief state
- decentralized control
- state space
- peer to peer
- decision problems
- planning under uncertainty
- cooperative
- partially observable
- qualitative and quantitative
- infinite horizon
- qualitative reasoning
- decision theoretic planning
- expected reward
- planning problems
- expert systems
- finite horizon
- policy iteration algorithm
- qualitative simulation
- markov decision process
- reward function
- distributed systems
- control policies
- partially observable markov decision process
- quantitative data
- approximate solutions
- qualitative models
- multiagent reinforcement learning
- decision theoretic
- influence diagrams
- decision making