Related Experiment Videos
Hierarchical global optimization of quasiseparable systems: application to Lennard-Jones clusters
1Novosibirsk State University, 630090 Novosibirsk, Russia. krivov@quentin.u-strasbg.fr
Physical Review. E, Statistical, Nonlinear, and Soft Matter Physics
|September 21, 2002
Abstract:
An unbiased hierarchical global optimization (GO) method for quasiseparable systems is presented. In such systems the coordinates of a set of nearby local minima approximately describe the coordinates of a much larger set of surrounding local minima. This allows one to reduce the original GO problem to a much simpler GO subproblem that uses the coordinates of the local minima to reduce the search space and simplify its landscape. The algorithm showed excellent performance in tests on "difficult cases" of Lennard-Jones (LJ) clusters. Putative global minima of LJ500 and LJ1000 are obtained.