Back To Index Previous Article Next Article Full Text


Statistica Sinica 5(1995), 575-591


CONTINGENTLY AND VIRTUALLY BALANCED

INCOMPLETE BLOCK DESIGNS AND THEIR

EFFICIENCIES UNDER VARIOUS OPTIMALITY CRITERIA


A. S. Hedayat, J. Stufken and W. G. Zhang


University of Illinois at Chicago, Iowa State University
and System Software Associates


Abstract: Even when a parameter set (vkλ) satisfies the necessary conditions for the existence of a Balanced Incomplete Block (BIB) design, the actual design may not exist or its existence may be unknown. We introduce two classes of designs, Contingently Balanced Incomplete Block (C-BIB) designs and Virtually Balanced Incomplete Block (V-BIB) designs, that may be considered in such cases. Both C-BIB and V-BIB designs are constructed from Unfinished Balanced Incomplete Block (U-BIB) designs, which can be constructed by a sequential search algorithm. Some V-BIB designs are shown to be highly efficient under A-, D-, and E-optimality criteria. Special attention is given to the parameter set (22, 8, 4) for which the existence of a BIB design is unknown. Highly efficient V-BIB designs exist for this parameter set. Also, C-BIB and V-BIB designs for (vkλ) may be used to construct BIB designs for parameter sets (vk) , where t>1 is an integer. This generalizes the well-known result that multiple copies of a BIB design form again a BIB design.



Key words and phrases: A-optimality, BIB(22, 33, 12, 8, 4), BIB designs, Block designs, D-optimality, E-optimality, Optimal designs, simple random sampling.



Back To Index Previous Article Next Article Full Text