A Two-Layer Framework for Joint Online Configuration Selection and Admission Control

Owen Shen, Haoran Xu, Yinyu Ye, Peter Glynn, Patrick Jaillet
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:111357-111392, 2026.

Abstract

We study online configuration selection with admission control problem, which arises in LLM serving, GPU scheduling, and revenue management. In a planning horizon with $T$ periods, we consider a two-layer framework for the decisions made within each time period. In the first layer, the decision maker selects one of the $K$ configurations (ex. quantization, parallelism, fare class) which induces distribution over the reward-resource pair of the incoming request. In the second layer, the decision maker observes the request and then decides whether to accept it or not. Benchmarking this framework requires care. We introduce a switching-aware fluid oracle that accounts for the value of mixing configurations over time, provably upper-bounding any online policy. We derive a max-min formulation for evaluating the benchmark, and we characterize saddle points of the max-min problem via primal-dual optimality conditions linking equilibrium, feasibility, and complementarity. This guides the design of SP-UCB–OLP algorithm, which solves an optimistic saddle point problem and achieves $\tilde{O}(\sqrt{KT})$ regret.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-shen26x, title = {A Two-Layer Framework for Joint Online Configuration Selection and Admission Control}, author = {Shen, Owen and Xu, Haoran and Ye, Yinyu and Glynn, Peter and Jaillet, Patrick}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {111357--111392}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/shen26x/shen26x.pdf}, url = {https://proceedings.mlr.press/v306/shen26x.html}, abstract = {We study online configuration selection with admission control problem, which arises in LLM serving, GPU scheduling, and revenue management. In a planning horizon with $T$ periods, we consider a two-layer framework for the decisions made within each time period. In the first layer, the decision maker selects one of the $K$ configurations (ex. quantization, parallelism, fare class) which induces distribution over the reward-resource pair of the incoming request. In the second layer, the decision maker observes the request and then decides whether to accept it or not. Benchmarking this framework requires care. We introduce a switching-aware fluid oracle that accounts for the value of mixing configurations over time, provably upper-bounding any online policy. We derive a max-min formulation for evaluating the benchmark, and we characterize saddle points of the max-min problem via primal-dual optimality conditions linking equilibrium, feasibility, and complementarity. This guides the design of SP-UCB–OLP algorithm, which solves an optimistic saddle point problem and achieves $\tilde{O}(\sqrt{KT})$ regret.} }
Endnote
%0 Conference Paper %T A Two-Layer Framework for Joint Online Configuration Selection and Admission Control %A Owen Shen %A Haoran Xu %A Yinyu Ye %A Peter Glynn %A Patrick Jaillet %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-shen26x %I PMLR %P 111357--111392 %U https://proceedings.mlr.press/v306/shen26x.html %V 306 %X We study online configuration selection with admission control problem, which arises in LLM serving, GPU scheduling, and revenue management. In a planning horizon with $T$ periods, we consider a two-layer framework for the decisions made within each time period. In the first layer, the decision maker selects one of the $K$ configurations (ex. quantization, parallelism, fare class) which induces distribution over the reward-resource pair of the incoming request. In the second layer, the decision maker observes the request and then decides whether to accept it or not. Benchmarking this framework requires care. We introduce a switching-aware fluid oracle that accounts for the value of mixing configurations over time, provably upper-bounding any online policy. We derive a max-min formulation for evaluating the benchmark, and we characterize saddle points of the max-min problem via primal-dual optimality conditions linking equilibrium, feasibility, and complementarity. This guides the design of SP-UCB–OLP algorithm, which solves an optimistic saddle point problem and achieves $\tilde{O}(\sqrt{KT})$ regret.
APA
Shen, O., Xu, H., Ye, Y., Glynn, P. & Jaillet, P.. (2026). A Two-Layer Framework for Joint Online Configuration Selection and Admission Control. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:111357-111392 Available from https://proceedings.mlr.press/v306/shen26x.html.

Related Material