E (χ2)-optimal mixed-level supersaturated designs

Mingyao Ai*, Kai-Tai Fang, Shuyuan He

*Corresponding author for this work

Research output: Contribution to journalJournal articlepeer-review

24 Citations (Scopus)


This paper introduces an optimality criterion for mixed-level supersaturated designs, called E (χ2), which covers E (s2) for two-level designs and ave (χ2) for symmetrical designs as two special cases. Some lower bounds of E (χ2) are obtained. Furthermore, it is shown that the E (χ2) criterion coincides with other existing criteria for nonregular mixed-level fractional factorial designs, which include minimum moment aberration, generalized minimum aberration, discrete uniformity and design efficiency. Several classes of E (χ2)-optimal mixed-level supersaturated designs from saturated orthogonal arrays are given.

Original languageEnglish
Pages (from-to)306-316
Number of pages11
JournalJournal of Statistical Planning and Inference
Issue number1
Publication statusPublished - 1 Jan 2007
Externally publishedYes

Scopus Subject Areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty
  • Applied Mathematics

User-Defined Keywords

  • Design efficiency
  • Discrete discrepancy
  • Generalized minimum aberration
  • Minimum moment aberration
  • Mixed-level
  • Orthogonal array
  • Supersaturated design


Dive into the research topics of 'E (χ2)-optimal mixed-level supersaturated designs'. Together they form a unique fingerprint.

Cite this