ASYMPTOTICALLY EFFICIENT ALLOCATION RULES FOR 2 BERNOULLI POPULATIONS

• 成果类型: Article
• 署名作者: LI, ZH; ZHANG, CH
• 署名单位: Rutgers University System; Rutgers University New Brunswick
• 刊物名称: JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY
• ISSN/ISSBN: 1369-7412
• 发表日期: 1992
• 页码: 609-616
• 关键词: bandit

摘要

Given two Bernoulli populations with unknown means, we consider the problem of sampling x1, x2, ... sequentially from these populations such as to maximize the expected value of the sum S(n) = x1 + x2 + ... + x(n) as n --> infinity. We obtain an asymptotically efficient rule. Simulation studies show that it performs well for moderate values of n.