Entropic Regularization of Markov Decision Processes
An optimal feedback controller for a given Markov decision process (MDP) can in principle be synthesized by value or policy iteration. However, if the system dynamics and the reward function are unknown, a learning agent must discover an optimal controller via direct interaction with the environment...
Main Authors: | Boris Belousov, Jan Peters |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2019-07-01
|
Series: | Entropy |
Subjects: | |
Online Access: | https://www.mdpi.com/1099-4300/21/7/674 |
Similar Items
-
A Relativistic Entropic Hamiltonian–Lagrangian Approach to the Entropy Production of Spiral Galaxies in Hyperbolic Spacetime
by: Michael C. Parker, et al.
Published: (2021-08-01) -
<i>α</i>-Geodesical Skew Divergence
by: Masanari Kimura, et al.
Published: (2021-04-01) -
Axiomatic Characterizations of Information Measures
by: Imre Csiszár
Published: (2008-09-01) -
Entropic Dynamics
by: Ariel Caticha
Published: (2015-09-01) -
Source Localization by Entropic Inference and Backward Renormalization Group Priors
by: Nestor Caticha
Published: (2015-04-01)