Statistical process control (SPC) is for monitoring sequential processes (e.g., production 
lines, Internet traffic, medical systems, social or economic status) to make sure that 
they work stably and satisfactorily. It is a major tool for quality control and management.
In the past 10--20 years, SPC has made great progress. Many new SPC methods have been 
developed for improving traditional SPC methods and for handling new SPC applications. This 
book aims to make a systematic description of both the traditional and recent SPC methods.

I started doing my research on SPC around 1998 after I joined the faculty of the
School of Statistics at the University of Minnesota. At that time, I was doing research on
jump regression analysis, which is about regression modeling when the underlying 
regression function has jumps. My senior colleague, Professor Doug Hawkins, published
a book on CUSUM charts that year, and he kindly gave me a free copy of his book. Before
I read the book, I had heard of CUSUM charts, but never had a chance to learn that subject 
systematically. The book makes a thorough description about the existing CUSUM charts. It 
has 10 chapters. Nine of them are on cases when a univariate quality
characteristic variable is of interest for process monitoring, and only one of them 
is on multivariate SPC, which is about cases with multiple quality characteristic variables 
being monitored. All multivariate SPC methods described in that chapter are based on the
assumption that the multiple quality characteristic variables follow a joint normal
distribution. In my opinion, there are two major limitations in the SPC research of that time.
First, in most applications, the quality of a product is affected by multiple characteristics
of the product (cf., Section 7.1 for a more detailed explanation). Therefore, the SPC
research should focus on multivariate cases, instead of univariate cases. Second, my extensive
consulting experience tells me that multivariate data would hardly follow a joint normal
distribution. Therefore, we should address the multivariate SPC problem without the
normality assumption. With encouragement from Doug, I started my own research on
multivariate SPC. After finding the fact that traditional multivariate SPC charts would
be unreliable in cases when the normality assumption was invalid (cf., Figures 9.1 and
9.2), I started to study existing statistical methods for describing multivariate non-normal
data and for transforming multivariate non-normal data to multivariate normal data. After
more than one year's study, I realized that existing statistical tools for describing and
handling multivariate non-normal data were limited, and the multivariate SPC problem
should be handled using certain statistical methods that were not based on the normality 
assumption. Since then, my co-authors and I have proposed antirank-based multivariate SPC 
charts that are appropriate to use in most applications and a general framework to handle 
the multivariate SPC problem using the log-linear model (cf., Sections 9.2 and 9.3). Besides
our research on multivariate SPC, my co-authors and I have also made some contributions
to univariate SPC, SPC based on change-point detection, and profile monitoring.

This book has 10 chapters. Chapter 1 describes briefly the concept of
quality, the early history of research on quality improvement, some
basic concepts on quality management, the role of SPC and other statistical methods 
in quality control and management, and the overall scope of the book. Chapter 2
describes some basic statistical concepts and methods that are useful in SPC. Chapters
3--5 make a systematic description of some traditional SPC charts, including the Shewhart, 
CUSUM, and EWMA charts. Some more recent control charts based on change-point detection
are described in Chapter 6. Some fundamental multivariate SPC charts under the 
normality assumption are described in Chapter 7. Then, Chapters 8 and 9 introduce 
some recent univariate and multivariate control charts designed for cases when the 
normality assumption is invalid. Control charts for profile monitoring are discussed in 
Chapter 10. At the end of each chapter, some exercises are provided for readers to
practice the methods described in the chapter. 
Computations involved in all examples in the book are accomplished using
the statistical software package R. Some basic R functions are introduced
in Appendix A, along with some R packages developed specifically for SPC analysis
and all R functions written by the author for the book. A list of
all datasets used in the book is given in Appendix B.

The mathematical and statistical levels required are intentionally low. Readers with 
some background in basic linear algebra, calculus through integration and differentiation, 
and an introductory level of statistics can understand most parts of the book without
much difficulty. 
For a given topic, some major methods and procedures are introduced in detail, and some more 
advanced or more technical material is briefly discussed in the section titled ``Some 
Discussions'' of each chapter. For some important methods, pseudo computer codes are 
given in the book. All R functions and datasets used in the book are posted 
on the author's web page for free download. 

This book can be used as a primary textbook for a one-semester course on statistical 
process control. This course should be appropriate for both undergraduate and graduate 
students from statistics, industrial engineering, systems engineering, management 
sciences, and other related disciplines that are concerned about process quality 
control. It can also be used as a supplemental textbook for courses on quality 
improvement and system management. SPC researchers from both universities and
industries should find this book useful because it includes many of the most recent research
results in various SPC research areas, including univariate and multivariate nonparametric 
SPC, SPC based on change-point detection, and profile monitoring. Quality control
practitioners in almost all industries should find this book useful as well, because 
many state-of-the-art SPC techniques are described in the book, their major advantages and 
limitations are discussed, and some practical guidelines about their implementations
are provided.

I am grateful to Doug Hawkins for introducing the SPC topic to me, for his 
guidance on my early SPC research, for his constant encouragement and help, and
for his numerous comments and suggestions during the course of my SPC research. 
I thank all my co-authors of SPC research, including Singdhansu Chatterjee, 
Doug Hawkins, Chang Wook Kang, Zhonghua Li, Zhaojun Wang, Jingnan Zhang, Jiujun Zhang,
and Changliang Zou, for their patience, stimulating discussions, and helpful comments
and suggestions. I am fortunate to have had Josh Wiltsie read the entire manuscript.
He provided a great amount of constructive comments and suggestions. Both Giovanna Capizzi
and Arthur Yeh provided detailed review reports about the book manuscript that greatly
improved the quality of the book. Part of the book manuscript was used as lecture notes
in my recent advanced topic course offered at the School of Statistics of University
of Minnesota in Fall 2012. Students from that class, especially Mr. Yicheng Kang, 
corrected a number of typos and mistakes in the manuscript. Dr. Changliang Zou kindly
helped me with the computation of the nonparametric EWMA chart described in Subsection 
8.2.3.

This book project took more than 3 years to finish. During that period of time, my wife,
Yan, gave me a great amount of support and help, by taking care of our two sons and
much household work as well. My two sons, Andrew and Alan, helped me in their own way by 
not interrupting me during my writing of the book manuscript at home and by keeping
my home office quiet. I thank all of them for their love and constant support.

Peihua Qiu 
Gainesville, Florida 
August 2013