With the scilab function solutionSetApproach(l,n,N,K) one can take a set with many members of length with many events. This can be repeated times. For every repetition the number of events will be multiplied by thereby increasing the number of events.Some empirical results can be seen below.
A first table (and diagram 3.16) shows 20 successive selections of n=32 individuals using N=64 until N=1280 events. While the mean value is mostly around 100% shows the half-distance between the maximum and the minimum of the mean values a decrease from about 19-17% to about 11-12% while increasing the number of events. Similarly shows the standard deviation in percentage a decrease from about 9.5 - 2.8. The 'real' mean values MEAN1 are centered around 0.031 with a standard deviation STD1 from 0.029 running down to 0.0008.
-->l=5, n=32, N=32, K=20, [MDX]=solutionSetApproach(l,n,N,K) MDX = N MEAN [%] DIST2 [%] STD MEAN1 STD1 -------------------------------------------------------------------- 64. 101.30208 17.1875 9.5991352 0.0316569 0.0029997 128. 99.699519 19.53125 8.8566371 0.0311561 0.0027677 192. 100. 15.364583 8.0806361 0.03125 0.0025252 256. 100.0625 13.476563 6.2994203 0.0312695 0.0019686 320. 99.951923 8.59375 5.5318223 0.0312350 0.0017287 384. 100.20461 7.9427083 4.1850025 0.0313139 0.0013078 448. 99.848214 9.8214286 4.419781 0.0312026 0.0013812 512. 100.49716 8.7890625 4.7191576 0.0314054 0.0014747 576. 99.807099 11.892361 5.4225295 0.0311897 0.0016945 640. 100.07813 6.71875 3.1863929 0.0312744 0.0009957 704. 99.784091 7.2443182 3.228669 0.0311825 0.0010090 768. 100.02325 8.1380208 3.4860207 0.0312573 0.0010894 832. 99.961367 8.8341346 3.4679514 0.0312379 0.0010837 896. 99.946121 5.859375 3.1022874 0.0312332 0.0009695 960. 99.934414 6.6666667 3.04064 0.0312295 0.0009502 1024. 100.10102 4.9804688 2.470903 0.0312816 0.0007722 1088. 100.17398 7.0772059 3.6741721 0.0313044 0.0011482 1152. 99.855324 5.3819444 2.952377 0.0312048 0.0009226 1216. 100. 5.7976974 2.8538176 0.03125 0.0008918 1280. 99.933036 5.0390625 2.8314077 0.0312291 0.0008848
In a last example the number of events has been increased until 12800. One can observe that the one-half difference becomes even more smaller approaching 1.7-2.8% deviance from the mean and the standard deviation runs from 3.3 to about 0.7 - 1.1 (cf. diagram 3.17).
-->l=5, n=32, N=1280, K=10, [MDX]=solutionSetApproach(l,n,N,K) MDX = N MEAN [%] DIST2 [%] STD --------------------------------------------- 1280. 100.18415 6.8359375 3.3885234 2560. 100.08594 3.2421875 1.7525262 3840. 99.927263 2.9817708 1.4970031 5120. 99.945313 3.4960938 1.7487093 6400. 99.972917 3.0703125 1.2020543 7680. 100. 2.0898438 1.0388803 8960. 100.01368 2.4497768 1.021788 10240. 100. 1.4501953 0.8440676 11520. 99.988799 1.7447917 0.7871516 12800. 100.01347 2.8632813 1.1348433
Gerd Doeben-Henisch 2012-03-31