{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/129242"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/129242","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Online learning algorithms design with applications to clinical trial, serving system and job scheduling problem","abstract":"Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo terms","abstract_html":"Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo terms","abstract_has_math":false,"creators":["Ruan, Yufei"],"institution":"University of Illinois Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Industrial Engineering","degree_department":null,"school":null,"contributors":["Zhou, Yuan","Srikant, Rayadurgam","Chen, Xin","Etesami, Rasoul"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-04-27","date_published":"2025-04-27","updated_at":"2026-07-22T22:25:04Z","subjects":["Multi-Armed Bandits","online learning"],"languages":["en","eng"],"rights":["Copyright 2025 Yufei Ruan"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/2142/129242","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Zhou, Yuan","Srikant, Rayadurgam","Chen, Xin","Etesami, Rasoul"]},{"key":"dc:creator","label":"Author","values":["Ruan, Yufei"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025-04-27","2025-05"]},{"key":"dc:type","label":"Dc Type","values":["text","Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Industrial Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Multi-Armed Bandits","online learning"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en","eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2025 Yufei Ruan"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/2142/129242"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo terms","The student, Yufei Ruan, accepted the attached license on 2025-04-24 at 00:35.","The student, Yufei Ruan, submitted this Dissertation for approval on 2025-04-24 at 00:45.","This Dissertation was approved for publication on 2025-04-27 at 14:27.","DSpace SAF Submission Ingestion Package generated from Vireo submission #21939 on 2025-10-19 at 18:10:52","This thesis presents novel algorithms and theoretical analyses for various online learning applications, including job scheduling, model selection in serving systems, and clinical trials. Each application is framed as a Multi-Armed Bandits (MAB) or Linear Bandits problem, tailored to its specific requirements. In the job scheduling application, we address the challenge of scheduling jobs on a set of machines in an online fashion with the overall quality of service as close as possible to an optimal offline benchmark. We design a variant of the Upper Confidence Bound (UCB) algorithm, achieving a logarithmic regret of O(ln T), which effectively balances exploration and exploitation in resource-constrained environments. For the serving system application, we develop a model selection algorithm to optimize the deployment of machine learning models across multiple servers in parallel. Framing this as a stochastic MAB problem, we propose UCB-based / Batch Elimination Framework algorithms for both online and offline settings. These approaches achieve regret bounds matching the full-information MAB regret of O(ln T). Simulations on an image classification task with the ImageNet dataset validate the algorithms’ efficiency and scalability. In the clinical trial application, motivated by the need to reduce the time and resource costs associated with treatment evaluation, we study the batched Linear Bandits problem. We prove that only O(log log T) batches are needed to achieve the minimax optimal regret of O( \\sqrt{dTmin{log K,d}}), even in the more restricted batch learning model. Along the way, this result proposes the distributional optimal design, a natural extension of the optimal experiment design, and provide a both statistically and computationally efficient learning algorithm for the problem, which may be of independent interest."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Online learning algorithms design with applications to clinical trial, serving system and job scheduling problem"]}]}],"canonical_facts":{"dc:contributor":["Zhou, Yuan","Srikant, Rayadurgam","Chen, Xin","Etesami, Rasoul"],"dc:creator":["Ruan, Yufei"],"dc:date":["2025-04-27","2025-05"],"dc:description":["Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo terms","The student, Yufei Ruan, accepted the attached license on 2025-04-24 at 00:35.","The student, Yufei Ruan, submitted this Dissertation for approval on 2025-04-24 at 00:45.","This Dissertation was approved for publication on 2025-04-27 at 14:27.","DSpace SAF Submission Ingestion Package generated from Vireo submission #21939 on 2025-10-19 at 18:10:52","This thesis presents novel algorithms and theoretical analyses for various online learning applications, including job scheduling, model selection in serving systems, and clinical trials. Each application is framed as a Multi-Armed Bandits (MAB) or Linear Bandits problem, tailored to its specific requirements. In the job scheduling application, we address the challenge of scheduling jobs on a set of machines in an online fashion with the overall quality of service as close as possible to an optimal offline benchmark. We design a variant of the Upper Confidence Bound (UCB) algorithm, achieving a logarithmic regret of O(ln T), which effectively balances exploration and exploitation in resource-constrained environments. For the serving system application, we develop a model selection algorithm to optimize the deployment of machine learning models across multiple servers in parallel. Framing this as a stochastic MAB problem, we propose UCB-based / Batch Elimination Framework algorithms for both online and offline settings. These approaches achieve regret bounds matching the full-information MAB regret of O(ln T). Simulations on an image classification task with the ImageNet dataset validate the algorithms’ efficiency and scalability. In the clinical trial application, motivated by the need to reduce the time and resource costs associated with treatment evaluation, we study the batched Linear Bandits problem. We prove that only O(log log T) batches are needed to achieve the minimax optimal regret of O( \\sqrt{dTmin{log K,d}}), even in the more restricted batch learning model. Along the way, this result proposes the distributional optimal design, a natural extension of the optimal experiment design, and provide a both statistically and computationally efficient learning algorithm for the problem, which may be of independent interest."],"dc:format":["application/pdf"],"dc:identifier":["https://hdl.handle.net/2142/129242"],"dc:language":["en","eng"],"dc:rights":["Copyright 2025 Yufei Ruan"],"dc:subject":["Multi-Armed Bandits","online learning"],"dc:title":["Online learning algorithms design with applications to clinical trial, serving system and job scheduling problem"],"dc:type":["text","Thesis"],"thesis:degree_discipline":["Industrial Engineering"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:04Z"}