Taguchi ideas

A major part of Taguchi s philosophy centers around the separation of effects caused by noise factors and effects caused by control factors. Control factors are variables that are under the control of the experimenter flow rate, pH, concentration, reactor temperature, etc. In contrast, noise factors are variables that are not under the control of the experimenter ambient temperature, ambient humidity, identity of process operator (Joe or Jane), source of raw material, etc. Control factors are sometimes called process factors noise factors are sometimes called environmental factors. 

The inner array is the large fractional factorial design shown in the center of 

The foldover design for the Plackett-Burman design in Table 14.9. 

An 11-factor Plackett-Burman design with foldover. 

Suppose viscosity, y, is the quality response of interest. Then at one of the four factor combinations in the inner array (i.e., in the control factor space), experiments can be carried out at the four factor combinations in the outer array (i.e., in the noise factor space). Frequently the outer array experiments are adventitious experiments in the sense that the experimenter has to wait until, say, the ambient humidity approaches the desired value then the appropriate source of raw material and the appropriate process operator can be brought in and the experiment can be carried out. 

---

A brief introduction into the Taguchi method is given in Chapter 4, whereas in Chapters 2 and 6 some of Taguchi s ideas are touched upon and discussed. The philosophy of Taguchi (build-in robustness) is present in Chapters 2,4, 6, 7 and 8. 

Taguchi s ideas can be separated into two fundamental concepts, which are ... 

In Taguchi methods, we define the quality of the product in terms of the deviation of some response parameter from its target or desirable value. The concept of quality and the idea behind the Taguchi philosophy is illustrated with the example shown in Figure 8.20. 

The inclusion of noise factors is one of the unique and important contributions of Taguchi. We illustrate the important role that noise factors can play and explore the relative advantages of the different paradigms in Section 2. The first paradigm, simple replication, leads largely to standard designs and analyses, and this approach is treated only briefly in Section 3 of this chapter. We discuss methods for analyzing experiments with noise factors in Section 4. Much of the research in this area has focused on what can be accomplished with no replication and these methods are presented in Section 5. We discuss some examples in Section 6 and summarize the ideas in Section 7. 

An important insight of Taguchi was that performance variation is often caused by natural variation in important input factors. Typical examples include the natural variation of a component part dimension about its nominal value or of a field condition at the time a product is used. Taguchi proposed that these variations should be expressly included as noise factors in the experiment. A simple example will clarify the idea and show how such an experiment differs from one with simple replication. Suppose an experiment is run on a product with a component part that has a nominal width of 3 mm. The parts are purchased from a supplier and have an average width equal to the nominal setting and a standard deviation of 0.05 mm. In a design with replication, the actual part used to build... 

We now consider factorial experiments in which there is no replication of design factor combinations and no use of noise factors. The idea of identifying dispersion effects in unreplicated factorials again has roots in the work of Taguchi. It was first studied in detail by Box and Meyer (1986) and has since attracted considerable interest and research. 

The best way to screen for dispersion effects is to include noise factors in the experiment and to exploit noise factor by control factor interactions. Experiments such as this will be most successful when the noise factors are indeed responsible for a large share of the outcome variation. Taguchi s idea of using noise factors to force controlled variation into experimental data is a striking and important contribution. 

In the meanwhile we explore some ways in which Taguchi s ideas may be applied using the experimental design strategy described in the rest of the book, and the kind of problems in pharmaceutical development that they are likely to help solve, in assuring reliable manufacture of pharmaceutical dosage forms. 

With respect to traditional thinking, the Taguchi approach, which later became quite popular in the West, introduced two new ideas ... 

Taguchi methods, and the other fractional factorials that we studied, have in common the idea of performing multivariate studies by means of orthogonal designs. Even if the researcher does not choose the ideal one, just by using them she will have more chances of succeeding than if she uses traditional univariate methods. 

Genichi Taguchi s methods have been widely known in industry for decades. The central idea of his methods is the quahty loss function and robust parameter design [42,43]. The quality loss function is used to estimate costs when the product or process characteristics are shifted from the target value. This is represented by the following equation ... 

Taguchi proposed that a measure of the loss to society, and to the customer, be a quadratic (square law) function of the difference between the target value and the actual value of the parameter of interest. This has become known as Taguchi s loss function. The goal posts idea just does not address this loss to society it does not penalize a product with a parameter that is not the target value. The loss function has been used as a measure of monetary loss associated with less than optimal performance as well as higher maintenance and repair costs. 

Taguchi has done quality control a service by his emphasis on a definition of quality and by his advocacy of robust design ideas. 

---