MATLAB: Simplifying a large nested for-loop

nested loops

I am trying to simplify a nested for-loop. Any suggestions would be highly appreciated!
The structure of the problem in its crudest form is the follows:
comb=zeros(1,N); %where N is a large number like 100
comb(1)=1;
for m2=1:N2 %where N2 is some predetermined number
comb(2)=IX(m2,2); %where IX is some pre-determined matrix
if consistent(comb) %where consistent(x) is a pre-specified fcn
for m3=1:N3
comb(3)=IX(m3,3);
if consistent(comb)
for m4=1:N4
comb(4)=IX(m4,4);
if consistent(comb)
...... %this needs to be continued in exactly the same fashion until I reach the (N-1)th nested loop
for mN=1:NN
comb(N)=IX(mN,N);
if consistent(comb)
tot_dist=min(tot_dist(comb),tot_dist) %where tot_dist(x) is a pre-specified fcn
else
end
end
......
else
end
end
else
end
end
else
end
end
The basic problem is conceptually the same as this: start from a vector comb=(1,0,…0). Select a number out of N2 numbers to fill the comb(2) if that selection satisfies some consistency condition. Do this for each of the remaining N-2 entries in comb, but the consistency condition is path dependent in the sense that your selection for comb(2) affects the consistency of a proposed selection for comb(3). Finally i need to select among all legitimate comb's the one that minimizes some total distance criterion.
Is there a way to handle this type of problems? Thank you very much in advance!
Yu

Best Answer

It sounds like a recursive problem to me.
What about something like
function comb = looper(comb, IX, n, N)
if n > length(comb)
return;
end
for m = 1:N(n)
comb(n)=IX(m,n);
if consistent(comb)
comb = looper(comb, IX, n+1, N);
end
end
end
You could call it with something like:
N = [N1, N2, N3, ..., Nn];
comb = looper(0, IX, 1, N);
tot_dist = min(tot_dist(comb), tot_dist);