Reputation: 20245
I am trying to generate all possible permutations of a set of elements. The order doesn't matter, and elements may be present multiple times. The number of elements in each permutation is equal to the total number of elements.
A basic recursive algorithm for computing permutations following the schema (as I am writing in C++, the code will look similar to it):
elems = [0, 1, .., n-1]; // n unique elements. numbers only exemplary.
current = []; // array of size n
perms(elems, current, 0); // initial call
perms(array elems, array current, int depth) {
if(depth == elems.size) print current;
else {
for(elem : elems) {
current[depth] = elem;
perms(elems, current, depth+1);
}
}
}
Would produce a large number of redundant sequences, e.g.:
0, 0, .., 0, 0
0, 0, .., 0, 1 // this
0, 0, .., 0, 2
. . . . .
. . . . .
0, 0, .., 0, n-1
0, 0, .., 1, 0 // is the same as this
. . . . . // many more redundant ones to follow
I tried to identify when exactly generating values can be skipped, but have so far not found nothing useful. I am sure I can find a way to hack around this, but I am also sure, that there is a rule behind this which I just haven't managed to see.
Edit: Possible solution+
elems = [0, 1, .., n-1]; // n unique elements. numbers only exemplary.
current = []; // array of size n
perms(elems, current, 0, 0); // initial call
perms(array elems, array current, int depth, int minimum) {
if(depth == elems.size) print current;
else {
for(int i=minimum; i<elems.size; i++) {
current[depth] = elems[i];
perms(elems, current, depth+1, i);
}
}
}
Upvotes: 2
Views: 129
Reputation: 316
Make your first position varies from 0 to n. Then make your second position to 1. Then make your first position varies from 1 to n. Then set second at 2 --> first from 2 to n and so on.
Upvotes: 2
Reputation: 49920
I believe one such rule is to have the elements in each sequence be in non-decreasing order (or non-increasing, if you prefer).
Upvotes: 2