10.The Design Theory and Its Application of Statistical Quality Control Charts for Complex Data
author:张玮光  Release time:2017-10-13   Browse times:31


    The research objective of this project lies in the interface between statistics and management science or industrial engineering, particularly on statistical quality control (SQC).

    Most of the traditional research on SQC focused on the quality of a process that can be adequately represented by the univariate or multivariate distribution of several specific quality characteristics. However, in many practical situations, say electronic products processes, the quality of a process is better characterized by a relationship between a response variable and one or more explanatory variables, i.e., profile monitoring, which offers great challenges to the conventional on-line monitoring methods. The problem of profile monitoring has been one of the most hot and difficult problems since its introduction in 2000. Most of the research in the literature focused on monitoring simple linear profiles before 2008. Two of our papers, published in the top journal of international industrial statistics Technometrics by the American Society for Quality and the American Statistical Association in 2008 and 2010 respectively, successfully solved the problem of monitoring complex profiles with arbitrary curves. The Editor commented our paper that "this is a valuable piece of work that provides a complete and solid solution for statistical monitoring of profiles". We published more than 10 papers along this research direction. Among them, one is the essential science indicators (ESI) highly cited papers, one is the 2nd of the most 10 highly cited papers during 2006-2008 in the flagship journal IIE Transactions by the International Institute of Industrial Engineers, two are the 3rd and 5th of the most 5 highly cite papers in Technometrics, and one is published in Technometrics as discussion paper, which is one of the 7 discussion papers in this journal during 2000-2010, and the researchers who discussed this paper are all world famous in the area of industrial statistics. Moreover, this paper was invited to be presented at the joint statistical meetings (JSM) in 2010, which is the largest scale statistical meeting in the world. It is a tradition that international top statistical journals publish discussion papers, but the proportion of discussion papers will not be more than 3% of the totally published papers. Furthermore, some of our research are cited in paragraphs more than 60 times in the book "Statistical Analysis of Profile Monitoring" published by the press Wiley. As for multivariate data streams and multi-stage process control, we made full use of the dimension reduction and variable selection methods in modern statistics to provide a unified framework for SQC and diagnosis. Many SQC research teams noticed and focused on our research, and made further research following ours. In the chapter "nonparametric (distribution-free) quality control charts" of the Encyclopedia of Statistical Sciences, the author Prof. Subha Chakraborti cited our research on nonparametric SQC in paragraphs.

      In summary, we published 49 science citation indexed (SCI) papers. Among them, two are in the journal Journal of American Statistical Association, nine are in the journal Technometrics, six are in the journal Journal of Quality Technology, three are in the journal IIE Transactions. All these papers are cited 351 times and the most one is even cited by 61 times. As members of the research team for this project, two of us won The National Excellent Doctoral Dissertation Award and Award nomination respectively, one of us won the Jiaqing Zhong Scholarship by the Chinese Mathematical Society, one of us won the Youth Science and Technology Award in Tianjin, one of us won the Outstanding Young Scholars of the First 10 Thousand Scholars Program by the Organization Department of the Central Committee of China.