"That way it makes it different every time I open matlab it is randomized hopefully differently."
Option 1: shuffle
rng('shuffle')
This command will choose a random seed based on the current time (sub-second resolution).
Option 2: randi
Another way to go about this is to save the seed to file and draw again if the seed has been previously chosen. Also, the seed can be any positive integer up to (2^32 - 1) so don't limit yourself with the number you arbitrarily chose.
This example assumes you have a mat file named 'seedlist.mat' that stores a vector named 'previousSeeds'.
%retrieve previous seeds
seedFile = load('seedlist.mat', 'previousSeeds');
previousSeeds = seedFile.previousSeeds;
% set rng, check for uniqueness, redo if needed
isUnique = false;
while ~isUnique
scurr = rng(randi(2^32-1)); %save seed & state; or scurr=rng('shuffle')
if ~ismember(previousSeeds, scurr.Seed)
isUnique = true;
end
end
% Store new seed to list on file
previousSeeds(end+1) = scurr.Seed;
save('seedlist.mat', 'previousSeeds', '-append')
I would also recommend you store a second variable that lists the date/time stamp that each seed was set. So you'd have a second vector that is the same size as previousSeeds. That way you have a log of what seeds were used so you can reproduce your data when needed.
One last word of caution, this affects everything that uses a random process in matlab so only do this when necessary.
