Heuristik, Genetischer Algorithmus und andere Zufälligkeiten in der Computerchemie

Computer Chemistry

Abstract

Heuristic procedures have been used increasingly in recent years in synthesis planning, chemometries and combinatorial chemistry. Applications of heuristic procedures have been presented in the chemical literature, but their use has been discussed only poorly. However, it has become evident that heuristic methods are not equally applicable to all problems. Indeed, the usefulness of the genetic algorithm can be deduced from the problem to be tackled: the most important prerequisite is that the quality does not playa significant role, i.e. a second best or even worse solution would be adequate and that the best solution is not necessary to solve the problem. Beyond that, the use of heuristics must be justified by the absence or high degree of complexity of non-heuristic methods.