Application of Genetic Algorithms to Structure Elucidation of Halogenated Alkanes Considering the Corresponding ~(13)C NMR Spectra

摘要

A new approach for structure elucidation using genetic algorithms is introduced. In analogy to the genetic programming paradigm developed by Koza, the new concept supports genetic operations on hierarchically coded chemical line notations. The implementation of this concept consists of 5 steps. In the first step, a start population of chemical compounds is randomly generated. As the second step, physical properties of each compound of the population are predicted. The third step is the comparison of each individual property with the observed property of an unknown compound, resulting in the calculation of the fitness value for each generated compound. Depending on the fitness values, the candidates for the next generation are selected by a spinning wheel procedure during the fourth step. In the last step, these candidates are rearranged by genetic mutation and crossover to form the next generation. Steps 2 to 5 of the described procedure are repeated until the spectrum of one candidate is almost equal to the spectrum of the unknown compound within acceptable tolerances. The introduced concept was verified for halogenated alkanes.