抄録
A random variable can be concentrated at only two sample points using two-point estimate method. However, this method is not practical for a large number of variables which are not depend on each other, because total number of sample points become 2' for n variable. This paper develops a simple procedure for reducing the sample points for several random variables. Through some numerical calculations, it is confirmed that the developed method can reduce the sample points keeping the accuracy for linear functions of several random variables.
