Genetic algorithms (GAs) have excelled in locating global minima but in many instances other low-lying minima are sought. A method is introduced here which has the analogue of predation in natural evolution, and which is linked to a variational theory on a metric space. The method is used to find the two most stable structures of small Morse, Al-n and C-n clusters.
|Number of pages||12|
|Journal||Match-Communications in mathematical and in computer chemistry|
|Publication status||Published - Oct 1998|