透過您的圖書館登入
IP:3.138.69.163
  • 期刊

Optimal Two-Level Fractional Factorial Designs with Pure Dispersion Factors

並列摘要


The problem of finding optimal designs when pure dispersion factors are present in tile class of regular single replicated two-level fractional factorial design of resolution Ⅲ and higher is studied. When a single dispersion factor is present, D-optimal and A-optimal designs depend on the number of length three words involving the dispersion factor in the defining relation for resolution Ⅲ designs. When two dispersion factors with equal dispersion main effect are present, D-optimal designs depend not only on the number of length three and length four words involving the dispersion factors in the defining relation but also on the values of the dispersion mean and main effects and the structures of the words. Tables are given to show how D-optimality ordering of designs changes when the values of the dispersion mean and main effects and the word structure change.

參考文獻


Box, G. E. P.,Meyer, D. C.(1986).Dispersion effects from fractional designs.Technometrics.28,19-27.
Brenneman, W. A.,Nair, V. N.(2001).Methods for identifying dispersion effects in unreplicated factorial experiment: A critical analysis and proposed strategies.Technometrics.43,388-405.
Chang, F. K.,Ting, C. P.(2009).Optimal two-level fractional factorial designs for location main effects with dispersion factors.Technical Report.(Technical Report).,::Department of Statistics, National Chengchi University.
Liao, C. T.,Iyer, H. K.(2000).Optimal 2n-p fractional factorial designs for dispersion effects under a location-dispersion model.Communications in Statistics-Theory and Methods.29,823-835.
Lin, M. Y.(2005).D-optimal regular 2n−p fractional factorial designs for location effects with single dispersion factor.Department of Agronomy, National Taiwan University.

延伸閱讀