This is the first report of a series of studies which deal which deal with the design problems of single variables sampling plans of the "point-of-control" type based on non-normal distributions. (A sampling plan of the "point-of-control" type named here means the plan characterized by two fundamental parameters, viz., the point of control p0.50 and the relative shope h0.50, as first introduced by Hamaker [4] in the theory of attributes sampling.) In this report we assume that the underlying distribution is of a Weibull type. In Section 3 the design problems for general case in question are solved by virtue of the Takagi's general theory [17] and those for Weibull case are solved in Section 4. Also, a series of conversion formulas between the fundamental parameters (p0.50, h0.50) for variables plans of the "point-of-control" type and the basic elements (p0,p1,α, β) for those of the "two-point specified" type is the plan characterized by two specified points, (p0,1-α) and (p1,β), on the OC curve, as seen in the Standards, e. g., JIS Z 9003 and 9004 based on normal distribution [11].) In Section 6 some examples which illustrate the use of conversion tables and sampling tables of Weibull plans are given. Finally, a single variables plan of the "point-of-control" type based on normal distribution viewed as a special case of the general solution given in Section 3 is discussed in Section 7.
View full abstract