Statistical Analysis For Evaluating Indices Of Product Process

Modern enterprises mostly manufacture their production in large scale. The conformance of the products and efficiency of product process are two important respects of business management. Process capability indices (PCIs) are important indexes to measure the conformance of the products, and stochastic frontier model functions are important tools for analyzing and measuring the efficiency of production. And there are academic value and actual meaning to carry on research to these two evaluation indexes.Many researches have already been paid to PCIs. Based on systemically summarize the existing achievement, this thesis chooses some difficult problems which have not been solved yet in this field to investigate further. The main innovative achievements are as follows: 1) To the multivariate process, in the occasion that both the specification and the distribution are an ellipse spheroid, for the first time, we put forward the multivariate process capability indices (MPCIs) which are suitable for various conditions such as centre, direction, form. The estimation of the distribution is studied by complicated academic deduce, and we get the maneuverable algorithms. 2) To the non-normal process, we completely solve the problem of the process controlling and capability assessing to log-normal distribution and Weibull distribution, with data transformation and equivalent quantization method, and we get perfect results when it is put into practice. 3) When process parameters are unknown, and the prior distributions are difficult to confirm, we construct empirical Bayesian estimation of univariate and multivariate PCIs depending on sample information and prove the convergence rates of estimation are up to 1. These achievements can obtain certain benefit to application in product process..There have been many researches to stochastic produce frontier functions too, but there are still many problems which are worth probing into. The chief subject...