Need help generating GA population

I am new to GA and there is something that not clear to me. I want to generate a population of k chromosomes with each chromosome length equal to Nt. Each gene in each chromosome is randomly either a '1' or '0'. For every '1' found in a chromosome i want to assign a random number between 1 and 15 for that index position. I need help generating this population.

Answer 1

[Star Strider]

Try this:

InitPop = randi([0 1], k, Nt);
InitPop(InitPop == 1) = randi([1 15], 1, nnz(InitPop));

This initially creates ‘InitPop’ as a matrix of [0 1] values, tne substitutes the 1 values with integers from 1 to 15.

[Farah Mahmood]

How do i retrieve the indices where new values are placed?

[Star Strider]

That requires a new line in my code:

InitPop = randi([0 1], k, Nt);
InitPop(Find1s) = randi([1 15], 1, nnz(InitPop));
[NewVals_r, NewVals_c] = find(InitPop);

The find call returns the indices of the non-zero elements in ‘InitPop’. Those are the row (‘NewVals_r’) and column (‘NewVals_c’) indices of the non-zero elements.

[Farah Mahmood]

Thanks, this gives exactly the rows and columns except i replace Find1s with InitPop==1.

I am treating each chromosome as a solution set so below is the first row (or first chromosome out of k)

0 0 0 5 0 6 5 0 0 2

so devices corresponding to positions 4,6,7,10 from the above chromosome are turned on

similarly 2 13 8 7 3 0 7 0 0 15 is the second chromosome and corresponding positions are 1,2,3,4,5,7,10

so on for the rest of k chromosomes.

I want display to read 'first subset solution' (shows 4,6,7,10)

'second subset solution' (shows 1,2,3,4,5,7,10) and so on and once GA finishes it shows the best subset.

Any ideas?

[Star Strider]

I mis-copied the code I intended to post. Should have been:

InitPop = randi([0 1], k, Nt);
InitPop(InitPop == 1) = randi([1 15], 1, nnz(InitPop));
[NewVals_r, NewVals_c] = find(InitPop);

To find the non-zero elements and present them for each row, do this before and after doing the optimisation:

[NewVals_r, NewVals_c] = find(InitPop);
RowIdx = accumarray(NewVals_r, NewVals_c, [], @(x){x});

for k1 = 1:size(InitPop,1)
    rowfmt = repmat('%2d ',1,numel(RowIdx{k1}));
    fprintf(['subset solution %4d (',rowfmt,')\n'],k1,sort(RowIdx{k1}))
end

This prints the indices of the non-zero elements.

Experiment with the format to get the result you want.

[Farah Mahmood]

Thanks a lot Star Strider, its doing something similar to what i wanted. Hopefully will work well with the rest of the code.

[Star Strider]

As always, my pleasure!