Towards Computationally Efficient Responsibility Attribution in Decentralized Partially Observable MDPs.
Stelios TriantafyllouGoran RadanovicPublished in: AAMAS (2023)
Keyphrases
- partially observable
- computationally efficient
- dec pomdps
- markov decision processes
- markov decision problems
- infinite horizon
- reinforcement learning
- state space
- decentralized control
- decision problems
- dynamical systems
- partial observability
- dynamic programming
- partial observations
- action models
- belief state
- computational complexity
- optimal policy
- reward function
- multi agent
- partially observable environments
- finite state
- fully observable
- partially observable markov decision processes
- average cost
- continuous state
- reinforcement learning algorithms
- average reward
- single agent
- probabilistic planning
- partially observable markov decision process
- np hard
- decision making