store value after multiple loops

Hi,

I have four different loops, from which I would like to store the results once an inside loops is finishing. The algorithm works like this:

Let xi, yi, xi, xi, sigmai be vectors with different numbers (values of parameters). What I am doing is I am looping over these parameter values to get some results based on an inner while loop with solving a particular algorithm. I, therefore, construct the following loops:

for n1=1:T
    x=xi(n1)

     for n2=1:T
         y=yi(n2)

         for n3=1:T
              z=zi(n)

                for n4=1:T

                    simga=sigmai(n4)
                    mu=z+y+x

                     while 1>1e-e  
                       results here that depend on the values of sigma and mu  
                    end

                end   
           end 
      end 
end

Can one suggest something practical in order to keep track and store the results in while loop for the different parameter values ?

Thanks

Answer 1

[Alexandra Harkai]

Storing results in a multidimensional array:

results = zeros(T, T, T, T);
% your code below
for n1=1:T
x=xi(n1);
  for n2=1:T
    y=yi(n2);
    for n3=1:T
      z=zi(n);
      for n4=1:T
        simga=sigmai(n4);
        mu=z+y+x;
        while 1>1e-e  
          results(n1, n2, n3, n4) = % results here that depend on the values of sigma and mu  
        end
      end
    end
  end
end

[msh]

Thank you for your answer, but there is a practical problem here. I cannot keep track of the different combinations of parameters that this results are produced. I was rather trying to see whether I can have a list(dataframe), or a cell array in which I will have one the matrix with the output of the results, another matrix with the combination of values, another matrices with other things that I want so save from the inside loop and so on.

[Guillaume]

It's really not clear what the problem is

"I cannot keep track of the different combinations of parameters that this results are produced" why not?

"I was rather trying to see whether I can have [...] a cell array in which I will have one the matrix [...], another matrix [...], another matrices with other things [...]" Sure. So why don't you create that cell array?

Also, do you actually need these 4 nested loops? Assuming that x, y, z are all row vectors. All possible mu can be created with:

mu = x + y(:) + permute(z, [1 3 2]);  %in R2016b
mu = bsxfun(@plus, x, bsxfun(@plus, y(:), permute(z, [1 3 2]))); %previous versions

[msh]

Guillaume, apologies for the confusion. I now worked out with a combination of your help and other sources what exactly I want to do.