Home
Resource
Asymptotic variance sensitive Markov decision processes
PDF
PDF format is widely accepted and good for printing.
Plug-in required
PDF-1(2.63 MB)
Citation & Export
View Usage Statistics
Staff View

Citation & Export
Hide

Simple citation

Benton, Andrew R.. Asymptotic variance sensitive Markov decision processes. Retrieved from https://doi.org/doi:10.7282/t3-4k9f-rw07

Export

Click here for information about Citation Management Tools at Rutgers.

Statistics
Hide

Description

TitleAsymptotic variance sensitive Markov decision processes
NameBenton, Andrew R. (author); Baykal-Gürsoy, Melike (chair); Coit, David (member); Gürbüzbalaban, Mert (member); Jafari, Mohsen A (member); Katehakis, Michael (member); Rutgers University; School of Graduate Studies
Date Created2022
Other Date2022-10 (degree)
SubjectIndustrial engineering, Markov processes, Statistical decision
Extent1 online resource (169 pages) : illustrations
DescriptionWe develop the asymptotic variance for Markov decision processes. Results are provided to express the asymptotic variance of a w-geometrically ergodic Markov reward process on a Borel space with randomized rewards. These results are then applied to problems minimizing the asymptotic variance over the space of randomized Markovian policies, which is generally a non-convex optimization problem due to the nonlinearity of the asymptotic variance. Dynamic programming algorithms and bilinear programming formulations are provided for variations of this task. The second part of this work develops reinforcement learning algorithms for this objective. Temporal-differences based estimators are developed, which are then applied to develop an actor-critic algorithm for mean-variance optimization.
NotePh.D.
NoteIncludes bibliographical references
Genretheses
Persistent URLhttps://doi.org/doi:10.7282/t3-4k9f-rw07
LanguageEnglish
CollectionSchool of Graduate Studies Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Version 8.5.5
Rutgers University Libraries - Copyright ©2024