品質管理相關知識(英文版)(ppt 24頁)

品質管理相關知識內容提要：
Screening Designs
A screening design is an experimental design whose purpose is to distinguish influential factors from non-influential factors, as efficiently as possible.
A fractional factorial design is a subset of the runs of a complete factorial design.
Example
For a certain process, there are 4 factors of potential interest. A 24 factorial design (16 runs) would have been a suitable experimental design.
However, resources of time and cost permit only 8 runs.
The design in last example is called a 24-1 fractional factorial design,
where 2 = number of levels for each factor
4 = number of factors
1 = degree of fractionation of the corresponding full factorial design
(in this case 2-1 = ? of the corresponding 24 factorial design)

The general notation for a 2-level fractional factorial design is 2k-p,
where 2 = number of levels for each factor
k = number of factors
p = degree of fractionation, or number of generators
Fractional factorial designs can be classified by their resolution.
Resolution III designs are those for which no individual factor is aliased with another individual factor, but for which at least one factor is aliased with a 2-factor interaction.
Resolution IV designs are those for which no individual factor is aliased with another individual factor or with any 2-factor interaction, but at least one 2-factor interaction is aliased with another 2-factor interaction, and at least one individual factor is aliased with a 3-factor interaction.
Resolution V designs are those for which no individual factor or 2-factor interaction is aliased with any other individual factor or 2-factor interaction, but at least one 2-factor interaction is aliased with a 3-factor interaction, and at least one individual factor is aliased with a 4-factor interaction.

..............................