target distribution

[Balduz]

Hello.

I would like to know if it is possible to extract the lines of a list considering a different re-distrtibuzione of parameters
I attach a file to make me better understand


what are the lines that can meet both the target distribution "a" together with the target distribution "b"

thank you very much for giving me a few minutes of your time

[MrShorty]

Is there a specific part of this analysis that you need help with? Do you have an algorithm already in mind, or do you need to develop the algorithm from the ground up? Without taking the time to think it all through thoroughly, here's how I think I would work this one out:

1) Spend some time with the algebra, to make sure I understand the different relationships. Things like n(1)+n(2)+n(3)=ntot [17+13+30=60] p(i)=n(i)/ntot [30/60=50%]. Given the p(i)'s and ntot, you should be able to compute all of the required n(i)'s and so on.
2) There will likely need to be some trial and error. The results of step (1) should hopefully help point the trial and error part of the algorithm to a good starting point.
3) Then work through the trial and error (maybe some linear programming) and what results. One like to hope that Solver could be useful here, but you will need to get a good set up for Solver to reliably converge on a suitable solution.