Monotonicity properties for two-action partially observable Markov decision processes on partially ordered spaces.
Erik MiehlingDemosthenis TeneketzisPublished in: Eur. J. Oper. Res. (2020)
Keyphrases
- partially ordered
- partially observable markov decision processes
- partial order
- partially observable domains
- partial ordering
- finite state
- totally ordered
- partial observations
- event calculus
- dynamical systems
- planning under uncertainty
- belief space
- dynamic programming
- decision problems
- optimal policy
- stochastic domains
- planning problems
- markov decision processes
- machine learning
- markov chain
- continuous action
- belief state
- infinite horizon
- reinforcement learning