osnaDocs: Reinforcement Learning with Recurrent Neural Networks

Reinforcement Learning with Recurrent Neural Networks

Please use this identifier to cite or link to this item:
https://osnadocs.ub.uni-osnabrueck.de/handle/urn:nbn:de:gbv:700-2008112111

Full metadata record

DC Field	Value	Language
dc.contributor.advisor	Prof. Dr. Martin Riedmiller
dc.creator	Schäfer, Anton Maximilian
dc.date.accessioned	2010-01-30T14:52:58Z
dc.date.available	2010-01-30T14:52:58Z
dc.date.issued	2008-11-20T16:23:42Z
dc.date.submitted	2008-11-20T16:23:42Z
dc.identifier.uri	https://osnadocs.ub.uni-osnabrueck.de/handle/urn:nbn:de:gbv:700-2008112111	-
dc.description.abstract	Controlling a high-dimensional dynamical system with continuous state and action spaces in a partially unknown environment like a gas turbine is a challenging problem. So far often hard coded rules based on experts´ knowledge and experience are used. Machine learning techniques, which comprise the field of reinforcement learning, are generally only applied to sub-problems. A reason for this is that most standard RL approaches still fail to produce satisfactory results in those complex environments. Besides, they are rarely data-efficient, a fact which is crucial for most real-world applications, where the available amount of data is limited. In this thesis recurrent neural reinforcement learning approaches to identify and control dynamical systems in discrete time are presented. They form a novel connection between recurrent neural networks (RNN) and reinforcement learning (RL) techniques. RNN are used as they allow for the identification of dynamical systems in form of high-dimensional, non-linear state space models. Also, they have shown to be very data-efficient. In addition, a proof is given for their universal approximation capability of open dynamical systems. Moreover, it is pointed out that they are, in contrast to an often cited statement, well able to capture long-term dependencies. As a first step towards reinforcement learning, it is shown that RNN can well map and reconstruct (partially observable) MDP. In the so-called hybrid RNN approach, the resulting inner state of the network is then used as a basis for standard RL algorithms. The further developed recurrent control neural network combines system identification and determination of an optimal policy in one network. In contrast to most RL methods, it determines the optimal policy directly without making use of a value function. The methods are tested on several standard benchmark problems. In addition, they are applied to different kinds of gas turbine simulations of industrial scale.	eng
dc.language.iso	eng
dc.subject	Optimal Control
dc.subject	Recurrent Neural Networks
dc.subject	Reinforcement Learning
dc.subject	State Space Reconstruction
dc.subject	System Identification
dc.subject.ddc	004 - Informatik	ger
dc.title	Reinforcement Learning with Recurrent Neural Networks	eng
dc.type	Dissertation oder Habilitation [doctoralThesis]	-
thesis.location	Osnabrück	-
thesis.institution	Universität	-
thesis.type	Dissertation [thesis.doctoral]	-
thesis.date	2008-10-31T12:00:00Z	-
elib.elibid	839	-
elib.marc.edt	jost	-
elib.dct.accessRights	a	-
elib.dct.created	2008-11-09T23:42:46Z	-
elib.dct.modified	2008-11-20T16:23:42Z	-
dc.contributor.referee	Dr. Hans-Georg Zimmermann
dc.subject.dnb	27 - Mathematik	ger
dc.subject.dnb	28 - Informatik, Datenverarbeitung	ger
dc.subject.ccs	I.2 - ARTIFICIAL INTELLIGENCE	eng
dc.subject.ccs	I.6 - SIMULATION AND MODELING	eng
vCard.ORG	FB6	ger
Appears in Collections:	FB06 - E-Dissertationen

Files in This Item:

File	Description	Size	Format
E-Diss839_thesis.pdf	Präsentationsformat	5,23 MB	Adobe PDF	E-Diss839_thesis.pdf View/Open

Show simple item record