Evolution, understood as the powerful interplay of reproduction, variation, and selection, represents an excellent tool for optimization in populations of simple and complex entities even under conditions where only limited information is available. The principle underlying this optimization heuristic was discovered and formulated already by Charles Darwin. In the nineteen-sixties evolutionary optimization became an intensively investigated topic. Experimental studies on evolution in the test-tube [1], mathematical analysis based on chemical reaction kinetics [2] as well as development of computer models and extensive simulations [3] provided detailed insight into the process on the molecular level: A repertoire of variants, called the molecular quasi-species, is created through error-prone reproduction and selection chooses among these variants those which have the highest reproductive success. A striking feature of the Darwinian optimization heuristic is its universal applicability. It is operative in ensembles of very simple systems like nucleic acid molecules and it is similarly successful in populations of highly complex entities, for example animal or human societies. The explanation of the success is straightforward: Selection at the population level is based exclusively on the mean numbers of fertile descendants in forthcoming generations, and therefore all mechanistic details of reproduction and variation are irrelevant for the survival of variants.