An integrated approach to solving influence diagrams and finite-horizon partially observable decision processes.
Eric A. HansenPublished in: Artif. Intell. (2021)
Keyphrases
- decision processes
- decision problems
- partially observable
- optimal policy
- influence diagrams
- finite horizon
- markov decision problems
- markov decision processes
- infinite horizon
- state space
- dynamic programming
- finite state
- reinforcement learning
- multistage
- sufficient conditions
- markov decision process
- long run
- average cost
- computational complexity
- initial state
- expected utility
- partially observable markov decision processes
- np hard
- decision making
- reward function
- single agent
- belief networks
- sensitivity analysis
- utility function
- training examples