MATLAB: Vectorization of while loop

Question

Hi all,

I have a small function, which generates random numbers in a specific interval. The function tests the generated values in respect to a certain bound and recalculates the random numbers again, if the bound is violated. I have to call this function several times (in this case 60 000 times), so this takes a lot of time (3/4 of the total time). I have read about vectorizing the code to improve performance. But I am totally helpless with this task. Your help would be greatly appreciated.

However, if someone has an idea to rewrite the while loop in a different way that would be great, as well.

Michi

Code:

meanValue = 10;
devValue = 1;
numValues = 10000;
presumption = 1;
for i = 1:60000
    arrayValues = RandomNums(meanValue, devValue, numValues, presumption);
end
function arrayValues = RandomNums (meanValue, devValue, numValues, presumption)
%% create rand nums
% normally distributed with meanValue +/- devValue
    numValues = round(numValues);
    arrayValues = devValue*randn(numValues,1) + meanValue; 
% check confidence-limes for each value and create new one if outside 
    for idx = 1:numel(arrayValues)
        while (arrayValues(idx) > (meanValue + presumption*devValue) || arrayValues(idx) < (meanValue - presumption*devValue))
            arrayValues(idx) = devValue*randn(1,1,'double') + meanValue;
        end
    end
end

Performance:

Answer 1

meanValue = 10;
devValue = 1;
numValues = 1000000;
presumption = 2;
  
tic
% Function from here https://www.mathworks.com/matlabcentral/fileexchange/23832-truncated-gaussian
arrayValues = meanValue + TruncatedGaussian(-devValue, presumption*[-1 1], [1 numValues]);
toc % Elapsed time is 0.055511 seconds for one billions random numbers.
% Check histogram
hist(arrayValues,100)

Histogram obtained