Secretaries with Advice br

• 成果类型: Article
• 署名作者: Dutting, Paul; Lattanzi, Silvio; Leme, Renato Paes; Vassilvitskii, Sergei
• 署名单位: Alphabet Inc.; Google Incorporated; Alphabet Inc.; Google Incorporated
• 刊物名称: MATHEMATICS OF OPERATIONS RESEARCH
• ISSN/ISSBN: 0364-765X
• 发表日期: 2024
• 页码: 856-879
• 关键词: markov decision-processes; combinatorial auctions; online; mechanism; algorithms; regret

摘要

The secretary problem is probably the purest model of decision making under uncertainty. In this paper, we ask which advice we can give the algorithm to improve its suc-cess probability. We propose a general model that unifies a broad range of problems: from the classic secretary problem with no advice to the variant where the quality of a secretary is drawn from a known distribution and the algorithm learns each candidate's quality on arrival, more modern versions of advice in the form of samples, and a machine-learning inspired model where a classifier gives us a noisy signal about whether the current secretary is the best on the market. Our main technique is a factor-revealing linear program (LP) that captures all of these problems. We use this LP formulation to gain structural insight into the optimal policy. Using tools from linear programming, we present a tight analysis of optimal algorithms for secretaries with samples, optimal algorithms when secretaries' qualities are drawn from a known distribution, and optimal algorithms for a new noisy binary advice model.