I need to optimize a model with 5 variables each of which cam be any valye from 1 to 20. The solution space, therefore, contains 3,200,000 combinations. I have a software which can give out a random sampling of this solution space. I need to know how many combinations can be considered as a representative sample of this solution space - e.g. can it be 1%?.
Aug 11th 2008, 05:54 AM
You have no Statistics backgound at all?
U.S. National elections tend to be +/- 3% with a sample well under 5,000. This compares to a voting population of over 300,000,000. A little arithmetic shows 5,000/300,000,000 = 0.00001666... or about 0.002%. I would say your 1% is a VERY LARGE sample, but it really depends on what you are doing and how much your audience might care.
Aug 12th 2008, 04:00 AM
thanks...i am trying to optimize a system...i truth it will have billions of possible input combinations so your reduction of the acceptable sample size percent is very much welcome!)))