Search for content and authors |
Cultural Differential Evolution: combined optimisation applied to structure solution from powder diffraction data |
Maryjane Tremayne , Samantha Y. Chong |
School of Chemistry, University of Birmingham, Edgbaston, Birmingham B152TT, United Kingdom |
Abstract |
Knowledge of the crystal structure of organic materials such as pharmaceuticals, pigments, agrochemicals, non-linear optics is essential if their solid-state properties are to be fully understood. However, when only polycrystalline samples of these materials are available, powder diffraction can be used as an alternative for structure determination. The most successful approach has been the development of direct space structure solution techniques in which a range of predicted models are compared with the experimental data using a global optimisation technique to locate the best crystal structure solution. This presentation will focus on our recent work on the development and application of the Cultural Differential Evolution (CDE) technique, combining the traditional biological dictates of mating, mutation and natural selection in the Differential Evolution algorithm (DE) with an approach that models human social behaviour or cultural selection. The DE method is a relatively new evolutionary algorithm that follows similar principles to conventional genetic algorithms, but differs in the execution of these processes. Our implementation of the Cultural Differential Evolution algorithm uses ‘cultural’ behaviour – in this case, the distribution in values of structural parameters in each generation – to guide and enhance the DE process through the use of dynamic DE boundaries (i.e. constraints) re-defined at each generation. This ‘dual inheritance system’ uses social factors to guide and speed-up evolution but natural biological selection to drive the optimisation process. Our results show an average 40% improvement in terms of the efficiency of a structure solution calculation, in all test cases, over the DE method alone, demonstrating a simple ‘real-world’ implementation of a combined cultural and biological evolutionary approach that has potential applications in many other optimisation problems in materials science. |
Legal notice |
|
Related papers |
Presentation: Invited at E-MRS Fall Meeting 2007, Symposium G, by Maryjane TremayneSee On-line Journal of E-MRS Fall Meeting 2007 Submitted: 2007-05-11 16:53 Revised: 2009-06-07 00:44 |