Customer Reviews

Statistical Control: By Monitoring and Feedback Adjustment

5 star:		(4)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

Professor George Box is a world renowned statistician, one of the greatest applied statisticians of the 20th Century. He has made fundamental contributions to the theory and practice of statistical methods. He is a wonderful lecturer and writer and can write on many levels. He has co-authored many classic books in statistics including texts on times series analysis and control, Bayesian methods, experimental designs, response surfaces and evolutionary operation. His work in design, response surface methods, time series and evolutionary operation were first motivated by his work in the chemical industry.
The text with Jenkins on time series analysis and control is a classic and had a major impact on the application of time series methods in business and industry. The systematic approach that they advocated was quickly referred to as the Box-Jenkins methodology. Although time series forecasting applications were well accepted by industry, the portion of the text demonstrating application to feedback control systems was largely ignored.

The 1980s and 1990s saw a resurgence of application of statistical process control in manufacturing. Basic Statistical Process Control (SPC) deals with control charts that assume independent identically distributed observations when "the system is under control". This is the basic theme for the Shewhart control charts and the cumulative sum charts that are standard in basic SPC. Hahn and Tucker at General Electric (among others) recognized that the approach could be generalized to dependent stationary processes using the Box-Jenkins methodology. In this text Box and Luceno describe this methodology and the recent results in the context of monitoring for process control and the incorporation of feedback adjustments to the process when the process goes out of control. Basically, they have taken the ideas on feedback control systems from the Box and Jenkins book and found applications to the control of manufacturing processes.

This is very useful stuff and is certainly worth knowing about. However if the reader is looking for a good introduction to standard SPC methods this is not the book he/she wants.

George Box and Alberto Luceno give you the best advice you need for Engineering Process Control (EPC) and SPC theory in this 300 page book. New methods, such as Cusum, and EWMA are also explained. Enough knowledge for a practician and more if you want in detail. They don't bore you by deriving equations you care less. I am an engineer in manufacturing and know when I need to study an equation or accept it as correct. In year 2000's, SPC and EPC will be emphasized in manufacturing even more than ever. Engineering process control has benefited from SPC in an exponentially growing rate. My motto is "SPC is the blood test of the manufacturing process control - S. Can". I recommend this book to all practicians of EPC.

The techniques described in this book have applications well beyond process or parts industries. I'm an engineer in mobile communications (an industry that desperately needs quality improvement) and I'm using SPC methods to eliminate choppy voice and dropped calls using feedback mechanisms between cell phones and base-stations. This book has provided me with a comprehensive framework for real-time network quality management. Don't be put off by its somewhat quaint language: this book gives real solutions to modern real-time signal processing problems.

This appears to be the first book to combine statistical process control (for discrete units) with automatic process control as used in the chemical process industry. As but one example, it provides a discrete measurement counterpart for proportional and integral (PI) control and includes damping factors to avoid overadjustment (tampering). It is NOT an introductory book to SPC, and some knowledge of SPC and time series analysis is helpful in understanding it.