Can i save random data in specific file in order to use them for different simulations?

Hello all! In my program code i have such random variables,they change for each iteration.Also there is some functions are depending on.These functions will generate results. In order to compare 3 approaches , i should use same results so i want to ask you if it is possible to save results for "100 iterations" and still use them for different approaches. hope that all clear, Thanks in advance.

[Walter Roberson]

perhaps use rng() to set the random number seed so they use the same random numbers?

[Maria]

Yes, I want to generate 100 different random variables and save them in specific matrices so I can use these variables for all scenarios in order to compare them using same values.

Answer 1

[Walter Roberson]

Yes you can do it. Whether it is a good idea to handle the situation this way is a different question.

rng(12345);
NumIter = 100;
MatrixSize = [73 19];
for iter = 1 : NumIter
    data = rand(MatrixSize);
    filename = sprintf('Matrix%03d', iter);
    save(filename, 'data');
end

[Maria]

@Walter Roberson Thank you for your suggestion but i want to know how to apply it in my situation,

for example i have a random matrix A = [5x10] already generated, i want to save "100" versions of A.

and how can i call them in different scripts?

[Walter Roberson]

NumIter = 100;

for iter = 1 : NumIter
    filename = sprintf('Matrix%03d.mat', iter);
    save(filename, 'A');
end

would store 100 exact copies of A in files named Matrix000.mat through Matrix100.mat .

I do not see the point in doing this, though.

It would make more sense to me if you were doing

NumIter = 100;
MatrixSize = [5 10];
for iter = 1 : NumIter
    A = rand(MatrixSize);
    filename = sprintf('Matrix%03d.mat', iter);
    save(filename, 'A');
end

which would store 100 different A matrices.

To use these in different scripts, somehow decide the filename to read from, and

datastruct = load(filename);
A = datastruct.A;

[Maria]

@Walter Roberson yes this is what i need exactly, but i want to read 100 versions of "A" at the same time. i mean that i wouldn't load one by one. I want to run my code for "100 iterations" at each iteration i will use the correspondant version of "A"

[Walter Roberson]

Well the easiest way to do that is

NumIter = 100;
MatrixSize = [5 10];
all_A = rand([MatrixSize NumIter]);
save('A.mat', 'all_A');

and then you would later

datastruct = load('A.mat');
all_A = datastruct.all_A;

then you would

NumIter = size(all_A,3);
for iter = 1 : NumIter
    A = all_A(:,:,iter);
    %now do whatever
end

[Maria]

@Walter Roberson i can't use this command line because A will be generated through a function and its size is changed each iteration (row size).

all_A = rand([MatrixSize NumIter]);

I tried to convert A in 3D matrix using this command line ,but it doesn't work

all_A = permute(reshape(A, 1, 2 ,3), [1 3 2]);

[Walter Roberson]

NumIter = 100;
all_A = cell(NumIter, 1);
for iter = 1 : NumIter
  A = YourFunction() ;
  all_A{iter} = A;
end
save('all_A.mat', 'all_A') ;

datastruct = load('all_A.mat');
all_A = datastruct.all_A;
NumIter = numel(all_A);

for iter = 1:Numiter
  A = all_A{iter};
  stuff
end

[Maria]

@Walter Roberson sorry for my late reply!

the part of "load" doesn't work,because the rest of my code is related to dimension of A so an error was displayed :(

[Walter Roberson]

Ummm, so rewrite your code in terms of size(A) ?

[Maria]

@Walter Roberson i will rewrite the code in cell may it works

[Maria]

@Walter Roberson thank you your for your help!