Ranking and Selection as Stochastic Control
Ranking and Selection as Stochastic Control
复制标题
DOI：
10.1109/tac.2018.2797188
复制
发表时间：
2017-10
影响因子：
6.8
通讯作者：
Yijie Peng;E. Chong;Chun-Hung Chen;M. Fu
中科院分区：
文献类型：
--
作者：
Yijie Peng;E. Chong;Chun-Hung Chen;M. Fu
在贝叶斯框架下，我们制定了完全序贯抽样和选择决策的统计排名和选择作为一个随机控制问题，并推导出相关的贝尔曼方程。利用价值函数近似，我们得到一个近似最优的分配策略。我们证明了这种策略不仅计算效率高，而且对于独立正态抽样分布具有一步前进性和渐近最优性。此外，建议的分配政策是很容易推广的近似动态规划范例。
Under a Bayesian framework, we formulate the fully sequential sampling and selection decision in statistical ranking and selection as a stochastic control problem, and derive the associated Bellman equation. Using a value function approximation, we derive an approximately optimal allocation policy. We show that this policy is not only computationally efficient but also possesses both one-step-ahead and asymptotic optimality for independent normal sampling distributions. Moreover, the proposed allocation policy is easily generalizable in the approximate dynamic programming paradigm.