### Table 4.1: Other Prede ned Quantities. 4.3 Basic MPI Operations The fteen functions represented in this section constitute a fairly reasonable subset of MPI. This subset is su cient enough to build many useful parallel applications. It may also be valuable in learning the fundamentals of the MPI model. These are the functions that were listed in Table 1.1.

### Table 2: Loss of fairness

2005

"... In PAGE 14: ... Thus, the network service provider can preserve the near optimal level of the desired revenue in each auction round. Remarkably, as shown in Table2 , the loss of fairness of BDC is lower than that of TA under the various wealth distributions, when bidders are allowed to drop out of an auction. The reason is that TA cannot prevent the loss of fairness caused by the higher bidders dropping out of an auction as a result of exceeding their TCL (i.... ..."

Cited by 3

### Table 2: Loss of fairness

2005

"... In PAGE 12: ... Thus, the network service provider can preserve the near optimal level of the desired revenue in each auction round. Remarkably, as shown in Table2 , the loss of fairness of BDC is lower than that of TA under the various wealth distributions, when bidders are allowed to drop out of an auction. The reason is that TA cannot prevent the loss of fairness caused by the higher bidders dropping out of an auction as a result of exceeding their TCL (i.... ..."

Cited by 2

### Table 4. We performed two ts, the rst being a power-law t and the second also allowing for a variation in the scalar spectral index (though in fact the underlying spectrum has none). The Figures and subsequent discussion use the former. The results for Models 1 and 2 contain no particular surprises. Although this is intended only to be indicative and certainly falls way short of the sophistication that can be brought into play on realistic data, the error bars are probably fairly reasonable. As expected, the tensor spectral index is the real stumbling block, but at least with these models one obtains a strong handle on A2 T , thus allowing a unique reconstruction. For these reconstructions, we nd that the lowest-order consistency equation Eq. (4.11) is indeed satis ed 0:108 0:013 = 2A2 T

"... In PAGE 38: ...f the potential. Reconstruction requires the in ation parameters in terms of observables. Relations between in ationary parameters and observables are given in Tables 3 and 4. A combination of information from Table 2 and Table4 results in Table 5, the observables needed to reconstruct a given derivative of the potential to a certain order. Although we know the information required for the next-to-next order given in Table 5, we don apos;t know the coe cients of the expansion.... In PAGE 67: ...bservable to the indicated order. (See Section 5.1.) parameter lowest-order next-order H A2 T A2 T , A2 S A2 T , A2 S A2 T , A2 S, n A2 T , A2 S, n A2 T , A2 S, n, dn=d ln k A2 T , A2 S, n, dn=d ln k ||{ Table4 : The in ation parameters may be expressed in terms of the observables, A2 T , A2 S, n, and dn=d ln k (see Section 5.1).... In PAGE 69: ... Showing the data in the form of the spectra is schematic; an analysis of true observations would directly t the amplitude and spectral index to measured quantities. Figure 3 The reconstructed potentials compared to the underlying one, from the data in Model 1 in Table4 . The dashed line shows the true underlying exponential potential.... ..."

### Table 1). Concerning SPEA, N was set to 4/5 and N to 1/4 of the population size given in Table 1, for reasons of fairness. Moreover, the domination pres- sure tdom, a parameter of NPGA, was determined experimentally. All NPGA simulations were carried out five times, each time using another value for tdom (5%, 10%, 15%, 20%, and 25% of the population size). At the end, the param- eter value which achieved the best results for the S measure was chosen per test problem (cf. Table 1).

### Table 4.3: Final Test Results The previous three solutions do not constitute a rigorous testing of the ac- curacy of the modelling and formulation; their aim is to show that a reasonable amount of consitency has been achieved with expected results in a fairly heuristic sense. The main testing will be performed by ICI sta , since only they have the true costs and bounds available.

### Table 5.6: Revised Comparison of Execution Cycles:Hardware fairly insigni cant with the SGA execution working on the 120-city problem since the execution time is large relative to the sampling period. With the adjustments made, the equations modeling the hardware ap- pear to be quite reasonable representations of the cycle costs for the GA operations and the overall execution times. The next chapter will present the equations mod- eling the software apos;s execution and provide results which indicate that the equation- based model is valid.

1996

### Table 5.13: Results on CELAR06 The Figures 5.18 and 5.19 show the decrease of the potential function, and the obtained solutions for CELAR06. The Figures 5.20, 5.21 and 5.22 show the performance of the algorithm for the preprocessed version of CELAR06. The solution of 4350 reported above, can be obtained by applying the 1{optimal algorithm to a solution of 5122, obtained by rounding scheme F1 after 10 iterations (appr. 300 seconds). We make the following observations: For the performance of the algorithm, the same remarks as in Section 5.3 apply. The rounding schemes F1 and F2 nd reasonable solutions fairly quick. Rounding scheme A generates in all cases the best solutions, but this takes substantial time. In the second run for the preprocessed version another local minimum than in the rst run is found. Unfortunately this yields higher cost than the local minimum found in the rst run. If we let the algorithm run, we may nd a better solution in the end. Since we are looking for good solutions obtained in reasonable time, no further results have been generated.

### Table 3. What are the key reasons for your company to be active in the field of safety, health and environment?

in Institute for Risk management and Safety analysis, Bergsprängargränd 2, S-116 35 Stockholm, Sweden.

"... In PAGE 6: ...isks, 4.03 for health hazards, and 3.94 for environmental risks. Most companies think that they have fairly good control of their risks. Reasons for SHE management A summary of the most important reasons for the companies to work with SHE issues can be found in Table3 , which also shows data obtained in the ACRONYM study of larger companies (Kok amp; van Steen, 1994a). The most important reason for working with SHE matters according to the SPASE study was quot;to comply with regulations quot;, followed by quot;publicity/image quot;, quot;pressure from authorities quot; and quot;pressure from employees quot;.... In PAGE 10: ... This was based on comparisons with the results from the ACRONYM study, which was focused on larger companies (Kok and van Steen, 1994a,b). The data can be found in Table3... ..."

### Table 3. What are the key reasons for your company to be active in the field of safety, health and environment?

in Preprint version of article published in Journal of Safety Research, vol. 31, no 2 pp. 71-80, 2000

"... In PAGE 6: ...isks, 4.03 for health hazards, and 3.94 for environmental risks. Most companies think that they have fairly good control of their risks. Reasons for SHE management A summary of the most important reasons for the companies to work with SHE issues can be found in Table3 , which also shows data obtained in the ACRONYM study of larger companies (Kok amp; van Steen, 1994a). The most important reason for working with SHE matters according to the SPASE study was quot;to comply with regulations quot;, followed by quot;publicity/image quot;, quot;pressure from authorities quot; and quot;pressure from employees quot;.... In PAGE 10: ... This was based on comparisons with the results from the ACRONYM study, which was focused on larger companies (Kok and van Steen, 1994a,b). The data can be found in Table3... ..."