Modeling and Planning with Macro-Actions in Decentralized POMDPs.
Christopher AmatoGeorge Dimitri KonidarisLeslie Pack KaelblingJonathan P. HowPublished in: J. Artif. Intell. Res. (2019)
Keyphrases
- macro actions
- markov decision processes
- reinforcement learning
- state space
- multi agent
- markov decision problems
- temporally extended
- partially observable
- planning under uncertainty
- finite state
- partially observable markov decision processes
- distributed constraint optimization
- orders of magnitude
- function approximation
- predictive model
- belief state
- decision processes
- learning algorithm
- temporal aspects
- dynamic programming
- action space
- reinforcement learning algorithms
- planning problems
- heuristic search
- optimal policy
- sufficient conditions
- probabilistic model