An Improved Measurement Utility at a Finite-Time Deadline
Abstract
Changing knowledge-gradient utility can alter the sequence well before asymptotic effects are visible. Finite-time analysis supplies a related question conditional on its reward model. Evaluation should preserve that model or explicitly test an adaptation, then compare complete independently checked runs at common total costs. Posterior utility is not measured policy superiority by itself. This focused synthesis develops an empirical comparison that preserves the terminal action, observation process, and total verification cost. Its proposed tests are prospective; no new simulation run, model-training result, or intervention effect is reported.
References
- L. Yang, S. Gao, and C. Ho, “Improving the Knowledge Gradient Algorithm,” Advances in Neural Information Processing Systems, pp. 61747–61758, 2023.
- Y. Li and S. Gao, “On the Finite-Time Performance of the Knowledge Gradient Algorithm,” Proceedings of the 39th International Conference on Machine Learning, pp. 12741–12764, 2022.
- Zhou, Yongxi, Lai Yun Choi, Jiaxi Wen, Wenbo Ye, and Junwei Yao. “How Much Can Repeated Running Buy? A Measured Variance Ceiling for LLM Code Evaluation.” arXiv, 2026, arXiv:2606.00920. https://arxiv.org/abs/2606.00920.
- Li, Y., Wang, H., Shang, X., Tang, X., Cao, Y., & Chen, X. (2026). BoostAPR: Boosting Automated Program Repair via Execution-Grounded Reinforcement Learning with Dual Reward Models. arXiv preprint arXiv:2605.09134.
- Chick, Stephen E., and Koichiro Inoue. “New Two-Stage and Sequential Procedures for Selecting the Best Simulated System.” Operations Research, vol. 49, no. 5, 2001, pp. 732–743. https://doi.org/10.1287/opre.49.5.732.10615.
- Frazier, Peter I., Warren B. Powell, and Savas Dayanik. “A Knowledge-Gradient Policy for Sequential Information Collection.” SIAM Journal on Control and Optimization, vol. 47, no. 5, 2008, pp. 2410–2439. https://doi.org/10.1137/070693424.
- Rasmussen, Carl Edward, and Christopher K. I. Williams. Gaussian Processes for Machine Learning. MIT Press, 2006. https://gaussianprocess.org/gpml/.
- Sacks, Jerome, William J. Welch, Toby J. Mitchell, and Henry P. Wynn. “Design and Analysis of Computer Experiments.” Statistical Science, vol. 4, no. 4, 1989, pp. 409–423. https://doi.org/10.1214/ss/1177012413.
- Bertsimas, Dimitris, and Melvyn Sim. “The Price of Robustness.” Operations Research, vol. 52, no. 1, 2004, pp. 35–53. https://doi.org/10.1287/opre.1030.0065.
- Campi, Marco C., and Simone Garatti. “The Exact Feasibility of Randomized Solutions of Uncertain Convex Programs.” SIAM Journal on Optimization, vol. 19, no. 3, 2008, pp. 1211–1230. https://doi.org/10.1137/07069821X.
- Glasserman, Paul. Monte Carlo Methods in Financial Engineering. Springer, 2004. https://doi.org/10.1007/978-0-387-21617-1.
- Ankenman, Bruce, Barry L. Nelson, and Jeremy Staum. “Stochastic Kriging for Simulation Metamodeling.” Operations Research, vol. 58, no. 2, 2010, pp. 371–382. https://doi.org/10.1287/opre.1090.0754.
Issue
Section
Articles