Picking Lognormal Random Numbers with a fixed sum?

Hello, i am trying to create a datasample out of the lognormal distribution, using lognormal random numbers.

Lets say i want to create 100 lognormal numbers with precalculated mean (here 17500) and varying shapes of the lognormal distribution

shapes=[0.4 0.57 0.74];

S=zeros([100 numel(shapes)]); %preallocate S Matrix

for kk=1:numel(shapes)

S = lognrnd(log(17500),shapes(kk),[kk,100]);

end

This gives me 100 lognormal random numbers in the datavalue range i want - but the problem is, for the 100 datapoints, the sum is fixed, here eg 1750000. I found this function / article

https://de.mathworks.com/matlabcentral/fileexchange/9700-random-vectors-with-fixed-sum

for uniformly distributed random numbers. But i didnt find any solution on normal or lognormal random numbers.

Thanks for your help and Best

Nico

Answer 1

[Nicolas Fischer]

Got it, nvm.