Why don't you use a matrix?
Let's say you have 10 timesteps (1 to 10),
ts = 1:10;
so now you'll need to track A's changes during every timestep. So create a 30x11 matrix and the first column being A's initial values.
A = zeros(30,11);
fill first column with initial values,
A(:,1) = rand(30,1);
now do something which changes this first column during every timestep,
for k=1:numel(ts)
A(:,k+1) = rand*A(:,1); %or even A(:,k+1) = rand*A(:,k);
end
now plot all changes,
plot(A)
you'll see 11 curves each for a timestep plus initial states.
