2024-03-29T15:50:14Zhttps://www.tdx.cat/oai/requestoai:www.tdx.cat:10803/73472024-03-15T10:57:51Zcom_10803_236col_10803_690269
TDX (Tesis Doctorals en Xarxa)
author
Ruíz Hernández, Diego
authoremail
diego.ruiz@upf.edu
authoremailshow
true
director
Glazebrook, Kevin
2011-04-12T16:33:06Z
2007-07-17
2007-04-13
9788469078266
http://www.tdx.cat/TDX-0717107-114628http://hdl.handle.net/10803/7347
B.39567-2007
In this Thesis, we first deploy Gittins index theory to establish the indexability of inter-alia general families of restless bandits that arise in problems of stochastic scheduling with switching penalties and machine maintenance. We also give formulae for the resulting indices. Numerical investigations testify the strong performance of the index heuristics.<br/><br/>The second class of problems concerns two families of Markov decision problems. The spinning plates problem concerns the optimal management of a portfolio of assets whose yields grow with investment but otherwise decline. In the model of asset exploitation called the squad system, the yield from an asset declines when it is utilised but will recover when the asset is at rest. Simply stated conditions are given which guarantee general indexability of the problem together with necessary and sufficient conditions for strict indexability. The index heuristics, which emerge from the analysis, are assessed numerically and found to perform strongly.
eng
Machine Maintenance
Switching Penalties
Indexability
Gittins Indices
Restless Bandits
Multi-armed Bandit Problems
Dynamic Programming
Stochastic Scheduling
Dynamic Allocation
Markov Decision Problems
Essays on indexability of stochastic sheduling and dynamic allocation problems
info:eu-repo/semantics/doctoralThesis info:eu-repo/semantics/publishedVersion
URL
https://www.tdx.cat/bitstream/10803/7347/1/tdrh.pdf
File
MD5
647b7cf495b8f06437c3ceaf6f11c310
1447798
application/pdf
tdrh.pdf
URL
https://www.tdx.cat/bitstream/10803/7347/2/tdrh.pdf.txt
File
MD5
a8d2cf3080113dc0eba28b506a00d480
373751
text/plain
tdrh.pdf.txt