MATLAB: Is there a way to generate pseudo-random integers directly in MATLAB 7.6 (R2008a)

MATLAB

I would like to generate a series of pseudo-random integers, but all the random-number generation algorithms in MATLAB either generate floating-point numbers, or threshold floating-point numbers to create integers. Is there a way to create these integers directly?

Best Answer

This enhancement has been incorporated in Release 2008b (R2008b). For previous product releases, read below for any possible workarounds:

For MATLAB 7.6 (R2008a) or earlier, if you have the Extended Symbolic Math Toolbox you can generate uniformly distributed pseudorandom integers using the MAPLE command in MATLAB. The following example illustrates how to generate a pseudorandom integer between 1 and 10:

myRandomInteger = maple('myProcedure:=rand(1 .. 10): myProcedure()')

Related Solutions

MATLAB: How to generate an unsorted permutation of integers in MATLAB 7.7 (R2008b)

This enhancement has been incorporated in Release 2009b (R2009b). For previous product releases, read below for any possible workarounds:

The RANDPERM function can be used to generate a random permutation of integers. For example, the following statement creates a random permutation of integers from 1 to 10:

randperm(10)

MATLAB: Uniformly distributed pseudo-random numbers

To do what you ask would require the result vector to have 145, not 144, elements, unless the vector is considered circular with the first element regarded as following the last element. In the following code you can have your choice since the last element in B always equals the first element. You can either leave B as is with its 145 elements or you can delete the last element, depending whether the vector is to be considered circular.

n = 12;
b = true;
while b
  A = zeros(n);
  B = zeros(n^2+1,1);
  j = ceil(n*rand);
  B(1) = j;
  for k = 1:n^2
    f = find(A(:,j)==0);
    m = length(f);
    if m == 0
      break;
    else
      i = f(ceil(m*rand));
      A(i,j) = k;
      j = i;
      B(k+1) = j;
    end
  end
  b = any(any(A==0));
end

This is not highly efficient code since it continues to repeat the while loop until it succeeds in completing the vector B with the requested property. This usually takes a number of trials. When it succeeds, the vector B will have one occurrence of each possible pairing of successive elements, and since the 'rand' function is used in this placement, B can be considered random in a sense.

Best Answer

Related Solutions

MATLAB: How to generate an unsorted permutation of integers in MATLAB 7.7 (R2008b)

MATLAB: Uniformly distributed pseudo-random numbers

Related Question