### Table 1: Numerical Results.

"... In PAGE 24: ... All the numerical results have been obtained by using the above strategy, with =5. Table1 summarise the numerical results, where we use the following abbreviations Prob: The problem number given in the Hock and Schittkowski collection [28]. Iter: The total number of iterations required to nd an optimum solution of the initial constrained problem (1).... ..."

### Table 2. Numerical results.

1995

"... In PAGE 7: ...Selected numerical results obtained are presented in Table2 . Values of the maximum and minimum overall empirical error rate are given for the subsequent stages in all knowledge acquisition processes which were conducted for the system based on the AQ15C learning algorithm (denoted AQ15C) and for the system using constructive induction-based learning (denoted AQ17- HCI).... In PAGE 7: ...The numerical results given in Table2 were used to construct learning curves for the average error rates for all combinations of learning and sampling, and they are given in Fig. 4.... ..."

Cited by 4

### Table 1. Numerical results.

1999

"... In PAGE 3: ... The differences between the listeners were small. Table1 gives a summary of the answers from the viewpoint of readers. All values are sums of average scores.... In PAGE 5: ... DISCUSSION First of all, there exist some notable differences between the speakers in the numerical results. As Table1 show, answers to questions about utterances spoken by the lecturer are the most accurate. It could also be clearly seen that the listeners had most difficulties when listening to the non-professional speaker.... ..."

Cited by 8

### Table 1. Numerical results

"... In PAGE 15: ...omputed by the planner. Here, the rolling bridge a simple system (4dof). However, the geometric complexity of the static scene ( 10:000 facets) and the large size of the movable object (which results in many RRT calls during the second stage) increase the overall cost of this connection stage. Table1 shows for the three examples numerical results of the performance of the algorithm. All examples were solved after less than ve minutes of com- putation.... ..."

### Table 1 Numerical results.

"... In PAGE 4: ...Hz clock and 1Mb of RAM memory, using OPL Studio 3.7.1 as the con- straint programming solver. We compared our results with the best over all those obtained by any of the MIP approaches in [5,8], according with the computation times presented in Table1 of [5] for the same 29 instances used in our study: 18 on complete graphs and 11 on sparse graph, with diameters equal to 4, 5, 6, 7, 9, and 10. Numerical results are presented on Table 1.... In PAGE 4: ... We compared our results with the best over all those obtained by any of the MIP approaches in [5,8], according with the computation times presented in Table 1 of [5] for the same 29 instances used in our study: 18 on complete graphs and 11 on sparse graph, with diameters equal to 4, 5, 6, 7, 9, and 10. Numerical results are presented on Table1 . For each instance, the first three columns give its number of vertices, its number of edges, and its max- imum diameter, respectively.... ..."

### Table 2 Numerical Results

"... In PAGE 9: ... This is due to the fact that the lower bound evaluation is not able to take into account tools. The results are listed in Table2 . It can be seen that in 74% of the cases the proposed solution is the optimal one, while in 96% of the cases the maximum error is lower than 2 setups.... In PAGE 9: ... It can be seen that in 74% of the cases the proposed solution is the optimal one, while in 96% of the cases the maximum error is lower than 2 setups. In 6 cases (marked with * in Table2 ) the problem resulted unfeasible and it was necessary to ask the upper level (Part Selection) for a new production plan with lower system saturation. Table 2 Numerical Results ... ..."

### Table 2. Numerical results

"... In PAGE 10: ... Thus they are able to provide much better starting points so that the ine ciencies we observe, are somewhat academic. In Table2 we collect the numerical results achieved. For each data set the least-squares algorithms DN2GB of Dennis et al.... In PAGE 10: ... However the numerical tests show that the solution of systems of nonlinear equations usually requires not more than 6 iterations with respect to the very small nal accuracy. In Table2 we list the name of the test case, the least-squares code DN2GB or DFNLP and the nal weighted residual. The subsequent two columns list the number of function and gradient evaluations of the outer least-squares algorithm until a termination condition is satis ed, i.... In PAGE 14: ...K. Schittkowski Table2 . (continued) problem code r? nf nrf i DN2GB .... ..."

### Table 1: Numerical results

"... In PAGE 6: ... Ad- ditional iterations merely slowed the algorithm and resulted in little improvement in solution quality. In Table1 , selected numerical results are shown.... In PAGE 8: ... It can also be seen that as the set size is reduced, the decrease in solution quality when the memetic algorithm is employed can be much less than the decrease observed when using the tabu search as the scheduling operator. From Table1 , the memetic algorithm clearly out- performs the tabu search algorithm in terms of so- lution quality, however tabu search can often find a result in a shorter time. In conclusion, with a small set size and little or no look ahead, almost a ten-fold decrease in run- time can be achieved with minimal increase in so- lution cost.... ..."