论文部分内容阅读
截止阀策略是解决序贯观察与选择问题的有效方法.在多次应用的问题情境下,通过减少截止阀值来降低失败的概率,进而实现平均收益的最大化,代价是减少了观察选项的数量,增加了决策的不确定性.本研究围绕降低失败的概率进行,从降低标竿而非减少截止阀值入手,采用均值标杆改进了最大值标竿的截止阀策略.基于仿真实验的结果显示,改进后的策略具有更好的平均收益和更低的失败概率,同时,还证明了20%就是均值标竿的近似最优解.
Globe valve strategy is an effective way to solve the problem of sequential observation and selection.In the case of multiple application problems, the probability of failure is reduced by reducing the cut-off threshold to maximize the average return, at the expense of reducing the number of observation options Quantity and increase the uncertainty of decision-making.This study focuses on reducing the probability of failure, starting with reducing the benchmark rather than reducing the cut-off threshold, and using the average benchmark to improve the strategy of the maximum benchmarking valve.Based on the simulation results It shows that the improved strategy has a better average return and a lower probability of failure, and at the same time, it proves that 20% is the approximate optimal solution of the mean standard.