Permissive Finite-State Controllers of POMDPs using Parameter Synthesis.
Sebastian JungesNils JansenRalf WimmerTim QuatmannLeonore WintererJoost-Pieter KatoenBernd BeckerPublished in: CoRR (2017)
Keyphrases
- reinforcement learning
- partially observable markov decision processes
- parameter values
- markov decision processes
- belief state
- data sets
- genetic algorithm
- distributed constraint optimization
- program synthesis
- finite state
- point based value iteration
- partially observable
- maximum likelihood
- evolutionary algorithm
- information systems
- search engine
- neural network