### Table 5: Best minimax design for the molding experiment

"... In PAGE 13: ...kWkXkj?1=8 jX0WkXj?1=8 = jX0 kWkXkj jX0WkXj ?1=8 : The matrices X and Xk are 256 8, with the rst 128 runs xed. The best design is displayed in Table5 . The minimax e ciencies of the ten best and ten worst designs are given in Table 6.... In PAGE 13: ... For the sake of comparison, we also augment the original design with 128 runs from the com- plementary half replicate and evaluate under the same criterion. In Table5 , we see that this design has quite a poor worst-case e ciency. Evidently, valuable information is lost from such a standard approach.... ..."

### Table 5 Kolmogorov-Smirnov with Lilliefors significance correction

### Table 5: Free design variables with upper and lower bounds.

"... In PAGE 4: ... Given this, the problem had to be significantly limited to a region of the design space with continuous variables only. Table5 shows the continuous variable used and their bounds. Table 6 shows the input variables held fixed and their assigned values.... ..."

### Table 1: Lower bounds on the number of physical links.

"... In PAGE 3: ... Most importantly we show in Theorem 5, that to embed rings of size 2 - N, a minimum of 4 2- N physical links are needed. A summary of these lower bound results is given in Table1 . In Section IV, these lower bounds are used to design physical topologies that can support rings of size with a minimal... ..."

### Table 1. ArcSim Variables and Ranges of Interest Design Variable Description Lower Bound Upper Bound

"... In PAGE 27: ...easured by the roll-over metric. The previous studies (Chen, et al., 1999) indicate that the roll- over metric has a highly nonlinear dependence on the control and noise variables, especially for different combinations of brake and steering levels. A description and the range of interest for each of the 14 input parameters are summarized in Table1 . All of the variables except brake_end have a range of +/- 20 % from their nominal values (i.... In PAGE 27: ... Five noise factors are chosen: three corresponding to the braking inputs, and two corresponding to the steering inputs. Their ranges are also listed in Table1 . The ranges of brake_start and brake are +/- 15% from their nominal values to avoid overlap of the two parameters, whereas the other parameter ranges are +/- 20% ... In PAGE 29: ....2. Experimental Set-Up for ArcSim Example The objective in this second example is to construct an accurate approximation model for the roll-over metric computed by ArcSim. There are a total of fourteen input variables as listed in Table1 , and the ranges of interest for each variable are also listed in Table 1. The factors and levels considered in this example are summarized as follows.... ..."

### Table 3. Engine Design Problem Identification Top-Level Design Specifications of the Propulsion System Design Design Parameters Lower Bound Upper Bound Baseli

"... In PAGE 21: ... A total of five engine cycle parameters are considered as the design parameters, which are modeled as the to-be-determined top-level engine design specifications. The upper and the lower bounds of these parameters are specified in Table3 a.... ..."

### Table 2. Upper and Lower Bounds of Design Variables

### Table 9: Test for Normality - Gender

in PRIOR PROGRAMMING EXPERIENCE AND THE INFLUENCE IT HAS ON STUDENTS' PERFORMANCE IN AN INTRODUCTORY

"... In PAGE 6: ...able 8: Mean and standard deviation of raw score for the four groups.................... 78 Table9 : Test for Normality - Gender.... In PAGE 88: ...05 (p = 0.200; see Table9 below), therefore the sample data can be assumed to be normally distributed. ... ..."

### Table 10: Test for Normality - Prior Programming Experience

in PRIOR PROGRAMMING EXPERIENCE AND THE INFLUENCE IT HAS ON STUDENTS' PERFORMANCE IN AN INTRODUCTORY

"... In PAGE 6: ...able 9: Test for Normality - Gender...................................................................... 78 Table10 : Test for Normality - Prior Programming Experience.... In PAGE 89: ...05 (p = 0.200; see Table10 below) therefore the sample data can be assumed to be normally distributed. ... ..."