Shao-Wei Cheng,Kenny Q. Ye
Abstract:Factorial designs have broad applications in agricultural, engineering and scientific studies. In constructing and studying properties of factorial designs, traditional design theory treats all factors as nominal. However, this is not appropriate for experiments that involve quantitative factors. For designs with quantitative factors, level permutation of one or more factors in a design matrix could result in different geometric structures, and, thus, different design properties. In this paper indicator functions are introduced to represent factorial designs. A polynomial form of indicator functions is used to characterize the geometric structure of those designs. Geometric isomorphism is defined for classifying designs with quantitative factors. Based on indicator functions, a new aberration criteria is proposed and some minimum aberration designs are presented.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is: in factorial designs involving quantitative factors, the traditional design theory treats all factors as nominal factors (i.e., unordered categorical variables), which is not appropriate. When the experiment involves quantitative factors, the permutation of factor levels may lead to different geometric structures, thus affecting the design properties.
Specifically, the paper aims to:
1. **Introduce the indicator function**: used to represent factorial designs and characterize the geometric structures of these designs in polynomial form.
2. **Define geometric isomorphism**: used to classify designs with quantitative factors.
3. **Propose a new deviation criterion**: based on the indicator function, propose a minimum - deviation criterion and give some examples of minimum - deviation designs.
### Key problems and solutions
- **Problem**: The traditional method cannot distinguish different geometric structures caused by the permutation of quantitative factor levels, resulting in differences in different design performances.
- **Solution**:
- Use the indicator function to represent the design and reveal its geometric structure in polynomial form.
- Define geometric isomorphism to classify factorial designs with quantitative factors.
- Propose a new deviation criterion (such as the β - word - length pattern) to more reasonably evaluate and select factorial designs with quantitative factors.
### Formula explanation
1. **Definition of the indicator function**:
\[
F_A(x)=\sum_{t \in T} b_t C_t(x)
\]
where \(C_t(x)\) is the orthogonal contrast basis function, \(b_t\) is the coefficient, and
\[
b_t=\frac{1}{N} \sum_{x \in A} C_t(x)
\]
2. **Conditions for geometric isomorphism**:
Designs \(A\) and \(B\) are geometrically isomorphic if and only if there exists a permutation \((i_1, i_2,\ldots, i_k)\) and a vector \((j_1, j_2,\ldots, j_k)\), where \(j_l\) is 0 or 1, such that
\[
b_{t_1 t_2\ldots t_k}=\left(\prod_{l = 1}^k (- 1)^{j_l t_{i_l}}\right)b'_{t_{i_1} t_{i_2}\ldots t_{i_k}}
\]
3. **New deviation criterion**:
\[
\beta_i(A)=\sum_{\|t\|_1 = i}\left(\frac{b_t}{b_0}\right)^2
\]
where \(\|t\|_1=\sum_{i = 1}^k t_i\) represents the total degree of the polynomial.
Through these methods, the paper provides a more reasonable way to process and evaluate factorial designs with quantitative factors, and solves the deficiencies of traditional methods in such designs.