Files in this item

FilesDescriptionFormat

application/pdf

application/pdf3130966.pdf (6MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:Theory and Application of Reward Shaping in Reinforcement Learning
Author(s):Laud, Adam Daniel
Doctoral Committee Chair(s):Gerald DeJong
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Computer Science
Abstract:We demonstrate our theory with two applications: a stochastic gridworld, and a bipedal walking control task. In all cases, the experiments uphold the analytical predictions; most notably that reducing the reward horizon implies faster learning. The bipedal walking task demonstrates that our reward shaping techniques allow a conventional reinforcement learning algorithm to find a good behavior efficiently despite a large state space with stochastic actions.
Issue Date:2004
Type:Text
Language:English
Description:97 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2004.
URI:http://hdl.handle.net/2142/81640
Other Identifier(s):(MiAaPQ)AAI3130966
Date Available in IDEALS:2015-09-25
Date Deposited:2004


This item appears in the following Collection(s)

Item Statistics