Optimal incomplete block designs

Alia Sajjad (University of Windsor, Canada)

When it comes to the best experimental designs statisticians need the ones that give them the estimates with minimum possible variance. Towards the quest of these designs different criteria can be used based upon the average variance, the maximum variance or the volume of the confidence region. This discussion gives the construction of optimal incomplete block designs with nearly minimal number of observations with respect to D-optimality (minimizing the volume of the confidence region) and A-optimality (minimizing the average variance) criteria. A unified approach towards the use of the associated graphs has been employed.