Cornell University
Sequential Resource Allocation Under Uncertainty: An Index Policy Approach
Abstract
dc:description.abstractWe consider a class of stochastic sequential allocation problems - restless multi-armed bandits (RMAB) with a finite horizon and multiple pulls per period. Leveraging the Lagrangian relaxation of the problem, we propose an index-based policy that uses the optimal Lagrange multipliers to index individual arms, and prove that the policy is asymptotically optimal as the number of arms tends to infinity. We also demonstrate numerically that this index-based policy outperforms state-of-the-art heuristics in several instances of RMAB. In addition, we study two other applications of sequential resource allocation problems which are extensions of the RMAB problem, and demonstrate how our index policy can be adapted to these settings.
Degree
thesis:*- Name thesis:degree_name
- Ph. D., Operations Research
- Level thesis:degree_level
- Doctor of Philosophy
- Discipline thesis:degree_discipline
- Operations Research
- Grantor
- Cornell University
- Year dc:date.issued
- 2017
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Hu, Weici
- Chair dc:contributor.chair
-
- Frazier, Peter
- Committee members dc:contributor.committeemember
-
- Topaloglu, Huseyin
- Joachims, Thorsten
Subjects
dc:subject × 5Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Dc Identifier Bibid
- 10361629
- Dc Identifier Other
-
ProQuest Submission ID: 10530
ProQuest Publication ID: 10618967