Resolution and interpreting of spectroscopic signalsof complex mixtures with the parallel genetic algortihms

Optimization is also an important calculation method used in analytical chemistry researches, as well as other research disciplines. Researchers used to optimize parameters and find relation between them by-trial-and-error method, before arise of artificial intelligence methods. As the number of parameters increases, it becames impossible to find a solution and even if there exist a solution it is generally not useful to conclude the problem. Genetic algorithm is one of the artificial intelligence techniques, recently used widely in computational sciences. As an optimization technique, genetic algorithm is described as an unusual example of research techniques.In this study the aim is to emphasize that genetic algorithm is a powerful technique for evaluating complex analytical signals and to show that it is a very useful tool for chemistry.Genetic algorithm is simply considered as finding correct answer in space and decreased problem solving duration by running more than one processors simultaneously(island model) by parallel computation. Different Island models were applied in this thesis and most effective one was selected for program. The developed software was tested with test functions in literature and compared with ?parallel and non-parallel? test functions.After these tests, the developed software was also tested by using the data obtained from our previous studies, Inonu University Central Research Lab. and chemistry databases. Standard error found during sollution of NMR, UV and IR data was compared with those from commercial software. It is found that the developed software gives better results compared to commercial software.As a result, it is concluded that since our results are near to experimental data the developed or written software can be safely used for analytical signals.