A heuristic approach for near optimal truncated sequential test of exponential distribution
A heuristic approach for near optimal truncated sequential test of exponential distribution
复制标题
DOI：
10.1080/07474946.2018.1554887
复制
发表时间：
2018-10
期刊：
影响因子：
--
通讯作者：
Sigui Hu;Honglei Wang
中科院分区：
文献类型：
--
作者：
Sigui Hu;Honglei Wang
摘要本文研究了指数分布的最优截尾序贯检验问题。讨论了最优截尾序贯检验的定义及相关概念。本文提出了一种求解近似最优截尾序贯试验的启发式方法--样本空间排序法（SSSM）及其求解步骤。结果表明，与国际标准IEC 61124（2012）和俄罗斯国家标准GOST R 27. 402（1995）提供的测试方案相比，本文提出的近最优测试方案能使测试误差概率与标称误差水平几乎相等，同时节省更多的综合期望测试次数和最大期望测试次数。
Abstract Optimal truncated sequential test of the exponential distribution is studied in this paper. Definitions and relative concepts of optimal truncated sequential test are discussed. A heuristic approach, sample space sorting method (SSSM) and the procedures are established to solve a near optimal truncated sequential test. Comparing with the test plans provided by international standard IEC 61124 (2012) and Russian national standard GOST R 27.402 (1995), the results show that the near optimal tests proposed in this paper can keep the error probabilities almost equivalence to the nominal error levels and save more synthetical expected test times and maximum expected test times simultaneously.