Monitoring Of Linear Profiles Using Change-Point Detection Method

With the development of economic globalization，the product quality has been paid moreattention, thus, techniques and methods to improve the quality of products have been extensivelyused and developed. Statistical Process Control (SPC for short) is the important part of productquality control investigation, it includes many efficient techniques for decreasing product qualitywave to keep the product quality steady, and control chart is one of them. With the wideapplication of compute and automation techniques, control chart gets a rapid development andachieves a good economic and social benefit. In most SPC applications, it is assumed that thequality of a process can be adequately represented by the distribution of a quality characteristic.However, in some situations, the quality of a process is better characterized and summarized by arelationship between the response variable and one or more explanatory variables. In particular,most of studies focus on the simple linear regression profiles.A control chart based on the sequential change-point detection model is proposed formonitoring the linear profiles. The structure of this dissertation is demonstrated as follows:In Chapter1, we introduce the outline of the background and history of SPC，including thedevelopment tendency and the hot issues in current SPC research.In Chapter2, we introduce the some relative knowledge and several basic control charts tohelp us understand this paper.In Chapter3, the simulated results show that our approach has good performance across therange of possible shifts. The simple diagnostic aids are also given to estimate the location of thechange and determine which of the parameters has changed.In Chapter4, we present an illustrative example.In Chapter5, we also investigate the effectiveness of these aids and give comparisons withthe traditional methods.At last，we summarize the main results and the technical approaches.The original and creative idea is this chart based on generalized likelihood ratio statisticscan detect a shift in either the intercept or the slope or the standard deviation by a single chartthat is different from the other control charts for linear profiles.