Application of a Genetic Algorithm in Structure Determination from Powder Diffraction Data
Benson M. Kariuki, Roy L. Johnston, Kenneth D. M. Harris, Katerina Psallidas, Shinbyoung Ahn and
Heliodoro Serrano-Gonzalez
School of Chemistry, University of Birmingham, Edgbaston,
Birmingham B15 2TT, U.K.
Article: MATCH - Communications in Mathematical and in Computer
Chemistry 1998, 38, 123-135.
Abstract
This paper describes the development of a Genetic Algorithm for
solving crystal structures directly from powder diffraction
data. A number of examples are given to illustrate the
application of this approach for the structure determination of
molecular crystals. Our approach adopts the direct-space
philosophy for structure solution from powder diffraction data,
in which trial structures are generated independently of the
experimental diffraction data and the quality of each trial
structure is assessed by comparing the calculated and
experimental powder diffraction patterns using the weighted
profile R-factor, R_wp. The use of the profile R-factor in this
approach implicitly takes care of the overlap of peaks in the
powder diffraction pattern, and avoids the need to extract the
intensities of individual reflections from the experimental
powder diffraction data. In our Genetic Algorithm, a population
of trial crystal structures is allowed to evolve subject to
well-defined rules governing mating, mutation and natural
selection, and the fitness of each structure is defined as a
function of R_wp. The aim is to find the minimum value of R_wp,
which corresponds to the best structure solution.
Postscript File