This is the first larger project that I ever worked on for myself :) I had a lot of fun with it :) Looking at it now, there are some things that I would do differently. But still. It runs for a few minutes, and it generates several hundred random initial structures. The resulting structures are good input structures for searching the potential energy surface of small clusters. When used as initial strutures for geometry optimization, the global minimum is always identified (or at least what seemed to be the lowest energy known minimum for the structures that I was working with).