"... In PAGE 18: ...1 Parameter Setting Besides population size AL and parent number AM, the strategy parameters B4DBBDBN BM BM BM BN DBAMB5, CRCR, CRCRD3DA, CRAR, and CSAR, connected to Equations (13), (14), (15), (16), and (17), respectively, have to be chosen.16 The default parameter settings are summarized in Table1 . In general, the selection related parameters AM, AL, and DBBDBN BM BM BM BN DBAM are comparatively uncritical and can be chosen in a wide range without disturbing the adaptation procedure.... In PAGE 18: ... By definition, all weights are greater than zero. In real world applications, the default settings from Table1 are good first guesses. Only for D2 BO BDBC does the default value yield AL BQ D2.... In PAGE 19: ... The optimal recombination weights depend on the function to be optimized, and it remains an open question whether the B4AMC1BN ALB5 or the B4AMCFBN ALB5 scheme performs better overall (using the default parameters accordingly). If overall simulation time does not substantially exceed, say, BED2 generations, CSAR should be chosen smaller than in Table1 , e.g.... In PAGE 23: ... 7 Simulation Results Four different evolution strategies are experimentally investigated. AF B4AMC1BN ALB5-CMA-ES, where the default parameter setting from Table1 is used apart from AL and AM as given below, and DBCX BP BD for all CX BP BDBN BM BM BM BN AM. To reduce the compu- tational effort, the update of BUB4CVB5 and BWB4CVB5 from BVB4CVB5 for Equations (13), (14), and (16) is done every D4D2 generations in all simulations.... In PAGE 25: ... This selection scheme is not recommended for the CMA-ES, where D2 BP BH (see Section 5.1, in particular Table1 and 2). The recommended B4BGCFBN BKB5-CMA-ES performs roughly ten times faster (compare also Figure 10).... ..."

"... In PAGE 11: ... For the CMA-ES, the default parameter setup suggested in [7] was adopted.This setup is illustrated in Table4 .The initial component-wise standard deviation of the mutation step, afii98460, was set to 1.... ..."

"... In PAGE 11: ... For the CMA-ES, the default parameter setup suggested in [7] was adopted.This setup is illustrated in Table4 .The initial component-wise standard deviation of the mutation step, afii98460, was set to 1.... ..."

"... In PAGE 6: ... We compare the results obtained by applying PSGES to those reported in [5, 1, 2] on the same benchmark. The results on the IEEE CEC 2005 benchmark for classic ES, PLES, LR-CMA-ES, IPOP-CMA-ES, as well as PSGES are listed in Table2 . The numbers are mean best objective function error values for 105 evaluations in 25 runs.... In PAGE 6: ... The values in the parentheses are the ranks for the performance. As we can see in Table2 , the results of PSGES is better than that of ES and PLES but worse than that of LR-CMA-ES and IPOP-CMA-ES for the five unimodal test functions (f1 to f5). Additionally, PSGES outperforms ES and PLES on 13 out of the 20 multimodal functions (f6 to f25), and LR-CMA-ES and IPOP-CMA-ES are ranked top 2.... ..."

"... In PAGE 8: ...able 11. Guidelines for prescription....................................................................................................... 75 Table12 .... In PAGE 32: ... These budgets also exist in Belgium, with indicative targets, in Greece for the main social insurance fund, in Italy and in Mexico. G3 Table12 . Fixed budgets, direct limitations of volume and expenditure 85.... In PAGE 44: ... Many OECD countries have developed and improved their prescription guidelines (Mitchell 1996) without however strictly linking them in a systematic way to financial incentives in the way that Fundholding did in England. OECD countries have often fixed arbitrary limits of consumption, either per day, per episode of care or per physician, but no clear picture has emerged of the results (See Table12 ). The OECD questionnaire reports explicit policies for the reduction of the volume of wasted drugs in a majority of countries.... ..."

"... In PAGE 4: ... The steps in the graphs are caused by the restarts that improve the performance on the noisy func- tion 4 and on the multi-modal functions 11 13, and 15 17. According to the requirements, Table1 reports CPU- time measurements, Table 2 gives the number of function evaluations to reach the success criterion (if successful), the success rate, and the success performances as de ned in Section 2. The objective function error values after 103, 104, 105 and n 104 function evaluations are presented in Table 3, 4, and 5.... ..."

"... In PAGE 4: ... The observed maximal nal population size is = 640; 448; 480, which means 26; 25; 25 times start = 10; 14; 15, for n = 10; 30; 50, respectively. According to the requirements, Table1 reports CPU- time measurements, Table 2 gives the number of function evaluations to reach the success criterion (if successful), the success rate, and the success performances as de ned in the previous section. The objective function error values after 103, 104, 105 and n 104 function evaluations are presented in Table 5, 6 and 7.... ..."

"... In PAGE 13: ... For each implementation, we selected the parameter values that resulted in the best accuracy. Table2 shows the parameter values and the results of the output modules. The CFinder algorithm is based on a clique percolation method.... ..."

"... In PAGE 5: ... That is, a higher precision search algorithm will move further down the ridge by taking smaller steps along the ridge direction [7]. Various measurements for both the 5-D Rana and Schwe- fel functions are listed in Table2 . For each function, CHC-10 and CHC-20 visit significantly more basins of attraction and have significantly higher mobility than either local search (LS-10 and LS-20) or CMA-ES (CMA-200 and CMA-500).... ..."