Input and Output Files
BLOCK.FOR
The program BLOCK.FOR computes the best possible designs for a blocked experiment. The search is over a finite grid of candidate points, which must be specified in the input. The default set of candidate points assumes a hypercubic design region and uses the points of a full factorial design as the set of candidate points. The following files can be used as an input to the program. The input files 1 to 4 use the default grid of candidates. The remaining input files contain a different set of candidates. Appendix A of Chapter 4 contains a concise overview of the program.
input1.prn
This input file can be used to generate D-optimal experiments with 3 blocks of 3 observations for the estimation of a full quadratic model in two quantitative variables. The D-optimal designs for this problem are displayed in Figure 4.2.
output1.prn
input2.prn
This input file is intended for the computation of a D-optimal design with 36 blocks of size 2 for quadratic regression on one explanatory variable. In the search, only the points -1, 0 and +1 are considered as candidates such that the resulting design is a minimum support design. Therefore, the optimal design does not depend on the degree of correlation (see Section 4.3.4).
output2.prn
input3.prn
This input file can be used for the computation of the best possible experiment with 3 factors and 2 blocks of 4 observations for the estimation of a linear model with two-factor interactions. The optimal design is the orthogonally blocked first order design displayed in Table 4.1. As shown in Section 4.3.1, this design is optimal for any degree of correlation.
output3.prn
input4.prn
This input file can be used for the computation of the best possible experiment with 5 factors and 32 observations for the estimation of a linear model with some two-factor interactions. The observations are arranged in 4 blocks of size 4 and 2 blocks of size 8. The optimal design is the orthogonally blocked first order design displayed in Table 4.2. As shown in Section 4.3.1, this design is optimal for any degree of correlation.
output4.prn
input5.prn
This input file is meant to compute the best possible design with 10 blocks of size 3 for the comparison of 5 treatments. The optimal design for this problem is the balanced incomplete block design given in Table 1.7.
output5.prn
input6.prn
This input file is meant to compute the best possible design with 10 blocks of size 3 for the comparison of 6 treatments. The optimal design for this problem is the balanced incomplete block design displayed in Table 4.3.
output6.prn
input7.prn
This input file is meant to compute the best possible design with 10 blocks of size 3 for the comparison of 9 treatments. The optimal design for this problem is the lattice design displayed in Table 1.8.
output7.prn
input8.prn
This file serves as input for the computation of a D-optimal design for the mixture experiment described in Section 4.3.4. The candidate points are the points from the second order simplex lattice design.
output8.prn