2013年10月15日 星期二

Pareto Distribution:Deming, Juran and Simon’s Rule and Worlds of Knowledge


Pareto Distribution:Deming, Juran and Simon’s Rule and Worlds of Knowledge 

( By Hanching CHUNG, 1998/11)

The readers of this website may like to know one interesting and important common or related subject which W. Edwards. Deming (1900-93), Joseph M. Juran  (1904-2008) and Herbert A. Simon (1916-2001) all spent some efforts and provided us some insights for understanding of the subject we are dealing with.

I am talking about the so-called Pareto Principle in quality or management field. Juran provided us with one important history of his discovery of his famous” 80/20 rule or Vital Few and Trivial Many rule “ in The Non-Pareto Principle; Mea Culpa. So as suggested by himself and agreed with many practical people like W. W. Scherkenbach that we should call it Juran Principle. Simon discovered independently the size distribution of different categories of the subject. Simon learned the principle from biology and rediscovered the rules and applied them to various economic fields like the evolving of the size of cities or business units. In fact, we can prove that all of them are talking with same size distribution phenomena. The extreme 95% dominating effects maybe the “ winner-take-it-all” effects. Deming and his friends used to tell that all the problems we encounter, 95% of them are system-related or 95% of the so-called improvement initiatives are futile because people don’t know the profound knowledge or competency..

The purpose of this article is trying to put them all together so that we have some insights of this important Pareto guideline. We’ll base our discussions on the idea proposed by David and Sarah Kerridge on the three worlds of knowledge.: Fact-World; Math-World; Mind-World. We’ll use the applications of so-called Pareto principle in the so-call quality improvement efforts(somebody put it right that most of these projects are mainly for the enhancement of productivity) by various products and service providing. We will demonstrate that in order to apply the Pareto rule effectively, we need to upgrade our knowledge in all the three worlds of knowledge.

According to one survey of the improvement tools used in the past decades of Japanese projects in quality improvement, Pareto diagram is one of the most popular one. Indeed, the distribution of the tools used followed the skew distribution we’ll discuss about in this article.

Juran put it right that many managerial activities follow Pareto distribution hence he use it as one of his main guiding principles in Managerial Breakthrough(1964). But in his work and the Japanese 7 quality-control tools, the original distribution-fitting and parameters estimation are replaced with simplified orders-ranking and simple calculations to pick up the Vital Few effects to be considered for improvement. This bring us to one issue that due to the simplification of Math-World, we may lose some valuable information hence we without insights we can derive from them which Simon and Ijiri can study the evolution and dynamics of the members of the studied population .

On the other hand, may be Juran and Deming’s insights are all for Mind-World of management of people and systems, their potentials are only started to be learned from most people. Simon’s approach may be limited in the further actions people can do to the system or industry. But may be in this age of pervasive computing, we need to make Math-World as completed as possible, that means to put original Pareto distribution information back for most practical people to make them as common tool of understanding of systems and variations. With this enhancements and better understanding of the insights we get from Deming, Juran and Simon, we may learn better and have a better way to know how to do the continual improvements.

We can do some reflection for the popularity of Pareto Principle in Quality Control Circle and other managerial activities. In certain way, current method is certain type of exploratory data analysis, it is good enough for a feeling of population studied. Maybe in many projects, it serve their purpose. Our point is we need to rethink the purpose of study and then apply the tools right. The trouble of most QCC or improvement team project is don’t know that system stability and apply tools blindly, few people can understand the relationship between Pareto Principle and control chart.

Knowledge is essential for application, and the Worlds of Knowledge concept remind us importance of system of profound knowledge.

Some notes on source of idea and. terms clarification.
J.M. Juran (1975) “ The Non-Pareto Principle; Mea Culpa” Quality Progress
J. M. Juran(1964,1994) “ Managerial Breakthrough” Mc-Graw Hill
David and Sarah Kerridge ( 1998 revised)“ Worlds of Knowledge” can be accessed through Deming Electronic Network.

The opinion of W.E. Deming were cited in The Fourth Generation Management by Brian Joiner(1994, Mc-Graw Hill). The Peter R. Scholtes’ rule were quoted in Deming’s The New Economics (1994, MIT CAES). Incidentally, we can compare one vehicle safety system study by W Mark. Crain(1980) that”…only a relatively small portion of highway accidents—some 2 to 6 percent—are conclusively attributable to mechanical(vehicle) defects.”( quoted from J. M. Juran’s A History of Managing for Quality(1994,ASQC Quality Press), p.574)

For Simon’s discovery journey, the best starting point may be his reflections in
Herbert A. Simon (1991)” Models of My Life” Basic Books, pp.371-75.

I also draw some information from Dr C. Martin Hinckley’s unpublished Doctor dissertation. A Global Conformance Quality Model(U): A New Strategic Tool for Minimizing Defects Caused by Variation, Error, and Complexity. I like to express my thanks to Dr. Hinckley.

Some clarifications of three worlds of knowledge.
Dr. Deming in his annex to The New Economics, used world 1,2,3 to describe relationship stage/maturity stages for suppliers’ partnership management.

David and Sarah Kerridge’s concept is very similar with Spinoza’s three type of Cogino ( Knowledge), see for example, The Cambridge Companion to Spinoza. In Buddhism or ancient Indian theory of knowledge, we can find similar thinking.

Jean Piaget’s three types of knowledge are physical; mathematical-logical, and social which is more congruent with common wisdom. In fact, another model of Worlds of Knowledge is W. W. Scherkenbach's model in his Deming's Road to Continual Improvement. For some thinking of this claasification, please refer to Kurt' Lewin(1936) Principles of Topological Psychology. He thought we need to discuss psychobiological organization Quasi-physical facts, Quasi-social facts and Quasi-conceptual facts. Piaget's thinking on children's "animistic" and true acquisition of knowledge can explain many Quasi-learning cases we encountered.

For the term of world 1, 2, 3, we should not ignore the most famous of late Sir Karl Popper’s works.. In fact we can learn a lot of interactions of different worlds from his “Intellectual Autobiography”, Chapter 4 of his Objective Knowledge and his Tanner’s Lecture ” On World 3”(1979, Michigan Quarterly).

The “Statistical Methods used in Presentations of QC Conference in Japan (1985-1996) “ table done by Ikuro Kusaba’s Keynote Speech “ Expectation for TQM” for 1997 CSQC Conference and the Asia Quality Symposium, pp. 1-6. It is illuminating to get a feeling of the essence of Japanese total 2034 presentations with the following application frequency:

graphic 1106; design of experiments 543; cause and effect diagram 524; histogram 442; tree diagrams 384;Preto diagram 380; scatter diagram 350; regression and correlation 210;QFD 189; control charts 136; others---

It is interesting to know that JUSE promoted NEW 7 QC TOOLS more than two decades but the result proved the basic cause-effect diagram and its related techniques are most popular. Most people don t bother to confirm the systems under study are in control states and Preto diagram is still among the vital ones.


沒有留言:

張貼留言