### Table 1. Values determining the D-optimal designs q k

"... In PAGE 5: ...nd k q ? 2. For q gt; 10 we did not compute the weights. However, it is clear that for k = 1; 0 and 0 are positive for all q. In Table1 values are given for q 8. Table 1.... In PAGE 6: ...misprints in Table1 of Fedorov (1972) on page 78. The table should read as follows.... In PAGE 17: ... We therefore have to calculate det(M( )) using the scaling ?1 xi 1 for the independent variables xi and the model (1). We obtain a new matrix X2 = XL, where X equals the matrix in Table1 1, and L = 0 B B B B B B B B B B B B @ 1 0 0 5=9 0 0 0 0 0 p5=3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 4=9 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 p5=3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 p5=3 1 C C C C C C C C C C C C A : (29) Therefore det(X0 2X2) = (det(L))2det(X0X) = 210 88=310 (30) and det(M( )) = 210=310 : This yields a D-e ciency of 82.4%.... In PAGE 25: ...5 The construction of designs with 16 experimental units using D-optimal designs The appropriate model to be used for the 4 4 4=16 design is the model described in (1) for k = q = 3. Using Table1 we nd that for k = q = 3 the D-optimal design consists of design points as given in Table 22. Table 22.... In PAGE 26: ... We nevertheless calculated the D-e ciency of the design in Table 23. For the design of Table 23 we obtain det((X0X)=16) = 1:566 10?4 : From Table1 and (8a) we obtain max det (M( )) = 0:000578313 : So, the D-e ciency of the design of Table 23 equals 87.8%.... ..."

### Table 2. Six Parameter D-Optimal Design

"... In PAGE 5: ... As a result, 91 design points (or parameter combi- nations) were eliminated due to excessive weight. The remaining 638 design combina- tions were used as a starting point for generat- ing an overdetermined D-optimal design with 45 design points or experiments ( Table2 ). 16 The number of experiments were chosen as 45 for comparison purposes since this many experiments were required by the CCD study.... In PAGE 5: ... 16 The number of experiments were chosen as 45 for comparison purposes since this many experiments were required by the CCD study. With this new design matrix ( Table2 ), the six parameters are studied at three levels (values) as represented in coded form by, -1, 0 and +1. As an example, a -1 for Fineness ratio corre- sponds to 4 (lower bound), a 0 corresponds to 5.... ..."

### Table 6. Average Kullback-Leibler divergences of \uniform quot; D-optimal U- designs and \non-uniform quot; D-optimal U-designs over 500 samples for the complete interaction Fourier model F (d; 1; ; 1; (R; H); 2) \uniform quot; design \non-uniform quot; design d n

"... In PAGE 11: ... For each sample and design the model was tted with the least square method and its predictive error was evaluated. Table6 lists the average Kullback-Leibler divergences of \uniform quot; based D-optimal U-designs and quot;non-uniform quot; ones over 500 samples for d = 4 and d = 5. It is apparent, from these two tables, that \uniform quot; D-optimal U-designs slightly outperform their competitors in most of the cases.... ..."

### Table 3: 6 Parameter D-Optimal Design Matrix (Unal et al, 1998a)

"... In PAGE 26: ...only the supersonic and hypersonic response surface models specifically include body flap deflection as one of the statistically significant random variables.) In the design of experiments for the weight function, forty-five multidisciplinary analyses were conducted according to the design matrix described in Table3 . For each row in the matrix, an analysis is run with design variable values given according to their corresponding column of normalized values.... ..."

### Table A4. D-Optimal versus Central Composite Design

in Launch Vehicle Design and Optimization Methods and Priority for the Advanced Engineering Environment

2003

### Table A6. D-Optimal Design Matrix

in Launch Vehicle Design and Optimization Methods and Priority for the Advanced Engineering Environment

2003

### Table 4.5: Runs that were made for the example in this article when ve response variables are used. Runs 1{8 come from an initial D-optimal design, runs 9{11 were runs in the optima of the (kriging) response surfaces of BHP, BHP1 and BHP2, based on the 8 rst runs, runs 14{16 were runs in the optima based on the 11 rst runs and using all 5 response variables. To improve the matches in runs 14{16, we made several iterations with new runs in the optimum of the response surface and a small D-optimal design for new runs near the optimum: Run 17 is a run in the optimum of the response surface based on the rst 16 runs, and runs 18{20 were designed around run 17. The next optimum was run 21, and runs 22{23 are the new local design. Finally, we made the runs 24 and 25 in the optima of the response surface based on 23 runs. The run that was used to generate the \history quot; is shown below (indicated with \H quot;).

### Table 3: Designs considered by DuMouchel and Jones (1994). Run BIRC A Design 2: D-optimal 10-point design for the primary model augmented with 2 center points. 1 0.0000 0.0000 0.3500 0.6000 0.0500

### Table 1: Sample sizes,n( max), n(!max) and n(N ) based on n( max), n(!max) and n(N ) respectively. The sample size is based on satisfying the curvature constraints based on evaluation at the individual support points of the prior and for the expectation over the prior, using the D optimal design for the Box-Tidwell example.

in Constrained Design Strategies for Improving Normal Approximations in Nonlinear Regression Problems

"... In PAGE 15: ... The maximum intrinsic curvature, maximum parameter e ects, and Beale apos;s measure N which combines intrinsic and parameter e ects curvature can be calculated for the di erent values in the support of 2. Using = 1 and the suggested guidelines, these constraints are converted into minimum samples sizes so that the constraints will be expected to be satis ed for each support point in the prior using the D optimal design ( Table1 ). The sample sizes... In PAGE 16: ... Replicates of this rounded optimal design were used to generate 100 simulated experiments at each of four sample sizes. The sample sizes used were n = 22, 33, 649, and 3234, suggested from Table1 . The sampling distributions for ^ 2, based on simulated experiments, under these four sample sizes are given in Figure 3.... ..."