<正> We consider a dam model of capacity Ⅴ units and make the following assumptions: (ⅰ) Let X(t) be the water input into the dam during the time interval (0, t], then {X(t):t≥0} is a Wiener process with nonn...<正> We consider a dam model of capacity Ⅴ units and make the following assumptions: (ⅰ) Let X(t) be the water input into the dam during the time interval (0, t], then {X(t):t≥0} is a Wiener process with nonnegative drift μ and variance parameter σ~2.展开更多
In this Paper, a parallel repairable model consisting of two units and one repairman isstudied. The working time and the repair time of the two units are all exponeotially distributed.Assume that one unit after repair...In this Paper, a parallel repairable model consisting of two units and one repairman isstudied. The working time and the repair time of the two units are all exponeotially distributed.Assume that one unit after repair will be 'as good as new', but the other one not. By introducingthe geometric process and using the method of supplementary variable, some importaDt reliabilityindlces are determined.展开更多
In this paper, we study the geometric process replacement model as follows:the successivesurvival times of the system form a nonincreasing geometric process while the consecutive repairtimes of the system constitute a...In this paper, we study the geometric process replacement model as follows:the successivesurvival times of the system form a nonincreasing geometric process while the consecutive repairtimes of the system constitute a non-decreasing geometric process, and the system is replacedat the time of the Nth failure after its installation or last replacement. Based on the long-runaverage cost per unit time, we determine the optimal replacement policy N,show the uniquenessof the policy N and discuss its monotonicity.展开更多
On testing for no effect of a predictor on response variable, some tests have been proposed in the literature. In this paper, one of them being a cusum test proposed by Buckley[1] is studied further. This test can be ...On testing for no effect of a predictor on response variable, some tests have been proposed in the literature. In this paper, one of them being a cusum test proposed by Buckley[1] is studied further. This test can be extended to treating the non-normal and time series cases. A Kolmogonov- Smirnov type test is suggested and studied too. The main observation is the link between the statistic and the Brownian bridge. Some small sample experiments are conducted to examine the power of the tests. In some non-normal cases the power is encouraging.展开更多
In this article, we study a model of a single variable sampling plan with Type I censoring.Assume that the quality of an item in a batch is measured by a random variable which follows aWeibull distribution W(λ,m), wi...In this article, we study a model of a single variable sampling plan with Type I censoring.Assume that the quality of an item in a batch is measured by a random variable which follows aWeibull distribution W(λ,m), with scale parameter A and shape parameter m having a gammadiscrete prior distribution or θ=1/λ and m having an inverse gamma-uniform prior distribution.The decision function is based on the Kaplan-Meter estimator. Then, the explicit expressions ofthe Bayes risk are derived. In addition, an algorithm is suggested so that an optimal samplingplan can be determined approximately after a finite number of searching steps.展开更多
文摘<正> We consider a dam model of capacity Ⅴ units and make the following assumptions: (ⅰ) Let X(t) be the water input into the dam during the time interval (0, t], then {X(t):t≥0} is a Wiener process with nonnegative drift μ and variance parameter σ~2.
文摘In this Paper, a parallel repairable model consisting of two units and one repairman isstudied. The working time and the repair time of the two units are all exponeotially distributed.Assume that one unit after repair will be 'as good as new', but the other one not. By introducingthe geometric process and using the method of supplementary variable, some importaDt reliabilityindlces are determined.
文摘In this paper, we study the geometric process replacement model as follows:the successivesurvival times of the system form a nonincreasing geometric process while the consecutive repairtimes of the system constitute a non-decreasing geometric process, and the system is replacedat the time of the Nth failure after its installation or last replacement. Based on the long-runaverage cost per unit time, we determine the optimal replacement policy N,show the uniquenessof the policy N and discuss its monotonicity.
基金the National Natural Sciences Foundation of China a Grant from the United College of Chinese University of Hong Kong
文摘On testing for no effect of a predictor on response variable, some tests have been proposed in the literature. In this paper, one of them being a cusum test proposed by Buckley[1] is studied further. This test can be extended to treating the non-normal and time series cases. A Kolmogonov- Smirnov type test is suggested and studied too. The main observation is the link between the statistic and the Brownian bridge. Some small sample experiments are conducted to examine the power of the tests. In some non-normal cases the power is encouraging.
文摘In this article, we study a model of a single variable sampling plan with Type I censoring.Assume that the quality of an item in a batch is measured by a random variable which follows aWeibull distribution W(λ,m), with scale parameter A and shape parameter m having a gammadiscrete prior distribution or θ=1/λ and m having an inverse gamma-uniform prior distribution.The decision function is based on the Kaplan-Meter estimator. Then, the explicit expressions ofthe Bayes risk are derived. In addition, an algorithm is suggested so that an optimal samplingplan can be determined approximately after a finite number of searching steps.
