Appendix A. DOE Schemes

Deterministic DOE Types

Table 1: Deterministic Design of Experiment

NameImageDescription
Full factorial  

  • Number of input parameter: k ≥ 1 (continuous and discrete)

  • Number of levels: p ≥ 2

  • Number of samples: n = pk

  • Linear terms

  • Mixed terms

  • Quadratic terms if p ≥ 3

Full Combinatorial  

This DOE generates all possible combinations of discrete input values

  • Number of input parameter: k ≥ 1

  • Number of discrete values pi

    for each dimension Xi

  • Number of samples:

  • Linear terms

  • Mixed terms

  • Quadratic terms if all pi ≥ 3

Star points  

  • Number of input parameter: k ≥ 1 (continuous and discrete)

  • Number of levels: p ≥ 1

  • Number of samples: n = 1 + 2pk

  • Linear terms

  • Quadratic terms

Central composite  

  • Number of input parameter: k ≥ 2 (continuous and discrete with ≥ 3 values)

  • Number of samples: n = 1 + 2k + 2k

  • Linear terms

  • Mixed terms

  • Quadratic terms

Box-Behnken  

  • Number of input parameter: k ≥ 3 (continuous and discrete with ≥ 3 values)

  • Number of samples: n = 2k(k −1) + 1

  • Linear terms

  • Mixed terms

  • Quadratic terms

D-optimal linear  

  • Number of input parameter: k ≥ 1 (continuous and discrete)

  • Number of samples:

  • Linear terms

D-optimal quadratic  

  • Number of input parameter: k ≥ 3 (continuous and discrete with ≥ 3 values)

  • Number of samples:

  • Linear terms

  • Mixed terms

  • Quadratic terms

D-optimal customized  

  • Number of input parameter: k ≥ 1 (continuous and discrete)

  • Number of samples:

  • Linear terms

Koshal linear  

  • Number of input parameter: k ≥ 1 (continuous and discrete)

  • Number of samples: n = 1 + k

  • Linear terms

Koshal quadratic  

  • Number of input parameter: k ≥ 1 (continuous and discrete with ≥ 3 values)

  • Number of samples:

  • Linear terms

  • Mixed terms

  • Quadratic terms