We review recent advances made by our group in finding optimized geometries of atomic clusters as well as in description of fission of charged small metal clusters. We base our approach to these problems on analysis of multidimensional potential energy surface. For the fusion process we have developed an effective scheme of adding new atoms to stable cluster isomers which provides good starting points for a global optimization procedure and thus allows one to obtain optimal geometries of larger clusters in an efficient way. We apply this algorithm to finding geometries of metal and noble gas clusters. For the fission process the analysis of the potential energy landscape calculated on the ab initio level of theory allowed us to obtain very detailed information on energetics and pathways of the different fission channels for the Na2+10 clusters.
provide the most useful information, given the assumed parameterization of the response probability surface. One drawback of global placement strategies is that if the form of the assumed threshold function is wrong, globalminimization may choose ( x, y ) locations that would be most informative if the