摘要:

In a production process it is necessary to specify tolerances for characteristics, such as dimensions within which the measured characteristic must fall in order for the part to be acceptable. Such tolerances may often be set informally, but a consideration of measurement error and the attendant risks of incorrect decisions can be helpful in establishing appropriate tolerances. We consider five measures of consumer's risk and evaluate them with respect to computational convenience, information requirements, rationale, and economic considerations.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979277

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

Using Grubb's measurement model, we establish an omnibus test for equality of bias and precision of measurement when two measuring devices are compared using paired data. The test is useful in controlling Type I error rates and increasing power in certain situations. Two examples of its application are given.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979278

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

The process capability indicesCpandCpkare widely used to provide unitless measures of process potential and performance. These indices do not adequately address the issue of process centering. An alternative definition ofCpadvocated by Taguchi addresses this issue directly. Later authors introduced the nameCpmfor the Taguchi index and examined statistical properties of an inefficient estimator under the assumption that the process mean coincides with the target value. The present paper presents statistical procedures based on the original Taguchi estimator which require no assumptions on the process mean.CpmandCpkare compared and contrasted to dispel the notion thatCpkmeasures process centering.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979279

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

Pareto Charts are a basic tool that derive their power from the Pareto principle. This work describes two statistical tests that complement the exploratory graphical analysis of the charts. The first test permits us to compare a given Pareto chart with another Pareto chart constructed over a different time period or process. The second test enables us to determine a relative ordering of sets of Pareto charts. The tests are presented in a step-by-step fashion with numerical examples.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979280

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

A number of models for the economic design of fraction nonconforming control charts have been presented. When multiple out-of-control states exist, a common assumption has been that at most a single assignable cause may affect the state of the production process during a specified or random interval. A model which does not constrain the number of shifts is presented and the effects of the“single shift”assumption are reviewed. Results indicate that the previous models' assumptions are valid for a significant range of model parameters. However, cases are presented for which some models are shown to be weakened by their limiting assumptions. In particular, the“single shift per process investigation”assumption is shown to exhibit a misleading relationship between the sampling interval and sample acceptance policy, such that an artificial degree of stability is conferred upon the production process when the expected interval between process investigations becomes large.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979281

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

We address the problem of choosing the most cost efficient set point for an automatic filling operation on a production line. Several researchers have examined this problem for the case where the filling operation is followed by 100% inspection. However, it is not always possible or economically justifiable to inspect every unit of product. In this paper we examine the case in which a sampling plan is used. There are two conditions examined, (1) when sampling results in destructive testing and (2) when the testing is nondestructive.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979282

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

Factors are provided for one-sided upper or lower, and two-sided prediction intervals for samples from a one-parameter exponential distribution. Some discussion regarding approximating these factors byfpercentiles with non-integer numerator degrees of freedom is also given.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979283

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

This paper deals with the problem of selecting the best manufacturing process (the one that has the largest process capability ratio index) from amongkavailable manufacturing processes. When the product quality characteristic data of each process follows a normal distribution, a modified likelihood ratio (MLR) selection rule is proposed. The sample size and the critical value, which are called for by this MLR rule, are computed by controlling the error probability and the probability of a correct selection. When the product quality characteristic data of each process follows a non-normal symmetric distribution, a simulation study, using the IMSL software, is used to study the robustness of this selection rule. The results suggest that the proposed rule is insensitive to certain symmetric non-normal (such as uniform and logistic) process distributions.

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979284

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

摘要:

Lenth (1989) proposes a quick and easy method for the analysis of unreplicated factorial and fractional factorial experiments. This article presents a FORTRAN computer program that uses Lenth's method to produce the simple graph, similar to the analysis-of-means, for determining the importance of the various effects (contrasts).

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979285

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor

ISSN:0022-4065    

DOI:10.1080/00224065.1991.11979286

出版商:Taylor&Francis    

年代:1991

数据来源: Taylor