Bonferroni-Free and Indifference-Zone-Flexible Sequential Elimination Procedures for Ranking and Selection
针对大规模模拟排序与选择问题，提出无需邦费罗尼校正且无差异区域参数的序贯淘汰程序，在保证概率保证的同时避免过度消耗计算资源。
The curse of dimensionality has long been one of the biggest challenges in solving large-scale simulation ranking and selection (R&S) problems. As the number of systems grows, existing approaches to R&S relying on the Bonferroni correction become increasingly conservative, rendering them overachieving in error control and consuming more computational resources than necessary. In “Bonferroni-Free and Indifference-Zone-Flexible Sequential Elimination Procedures for Ranking and Selection,” Wang, Wan, and Chen develop Bonferroni-free and indifference-zone-optional ranking and selection procedures to deliver the prescribed probabilistic guarantee without overshooting. Their approach is to conduct always valid and fully sequential hypothesis tests that enable continuous monitoring of each candidate system and control the probability of correct selection. In addition, the indifference-zone parameter becomes dispensable in their procedures; however, when provided appropriately, it could improve the procedures’ computational and statistical efficiency.