The Knowledge-Gradient Policy for Correlated Normal Beliefs 论文

2009INFORMS journal on computing引用 434
Machine Learning and AlgorithmsReservoir Engineering and Simulation MethodsGaussian Processes and Bayesian Inference

详细信息

发表期刊/会议
INFORMS journal on computing
发表日期
2009-05-20
发表年份
2009

关键词

Machine Learning and AlgorithmsReservoir Engineering and Simulation MethodsGaussian Processes and Bayesian Inference

摘要

We consider a Bayesian ranking and selection problem with independent normal rewards and a correlated multivariate normal belief on the mean values of these rewards. Because this formulation of the ranking and selection problem models dependence between alternatives' mean values, algorithms may use this dependence to perform efficiently even when the number of alternatives is very large. We propose a fully sequential sampling policy called the knowledge-gradient policy, which is provably optimal in some special cases and has bounded suboptimality in all others. We then demonstrate how this policy may be applied to efficiently maximize a continuous function on a continuous domain while constrained to a fixed number of noisy measurements.