Previous | Table of Contents | Next |
Sometimes the number of experiments required for a full factorial design is too large. This may happen if either the number of factors or their levels is large. It may not be possible to use a full factorial design due to the expense or the time required. In such cases, one can use only a fraction of the full factorial design. Later in Chapter 19 we discuss the procedure to design a specific class of fractional factorial designs. Here, we simply give an example.
TABLE 16.3 A Sample Fractional Factorial Design | ||||
---|---|---|---|---|
Experiment Number | CPU | Memory Level | Workload Type | Educational Level |
1 | 68000 | 512K | Managerial | High school |
2 | 68000 | 2M | Scientific | Postgraduate |
3 | 68000 | 8M | Secretarial | College |
4 | Z80 | 512K | Scientific | College |
5 | Z9D | 2M | Secretarial | High school |
6 | Z80 | 8M | Managerial | Postgraduate |
7 | 8086 | 512K | Secretarial | Postgraduate |
8 | 8086 | 2M | Managerial | College |
9 | 8086 | 8M | Scientific | High school |
For every advantage there is a corresponding disadvantage. Fractional factorial designs save time and expense when compared to full factorial designs. However, the information obtained from a fractional factorial design is less than that obtained from a full factorial design. For example, it may not be possible to get interactions among all factors. On the other hand, if some of the interactions are known to be negligible, this may not be considered a problem and the time and expense of a full factorial design may not be justified.
In the remainder of this book, we restrict our discussion to the full factorial and fractional factorial designs.
Previous | Table of Contents | Next |