Accelerated cluster structure search using electron diffraction data in a genetic algorithm

We present a new method to accelerate the convergence behavior of a genetic algorithm (GA) for global cluster structure search by adding experimentally determined gas phase electron diffraction information in terms of a profile factor as additional criterion for evaluating trial structures into the objective function of the GA. This algorithm has been applied on two test systems: in case of B 18 + the superior ability of locating the global minimum in a narrow side potential energy funnel is shown. In the second test case, Au 13 - , the robustness of the modified GA against systematic errors in the energy calculation is demonstrated.