Application of genetic algorithm for extraction of the parameters from powder EPR spectra

The application of the stochastic genetic algorithm in tandem with the deterministic Powell method to automated extraction of the magnetic parameters from powder EPR spectra was described. The efficiency and robustness of such hybrid approach were investigated as a function of the uncertainty range of parameters, using simulated data sets. The discussed results demonstrate superior performance of the hybrid genetic algorithm in fitting of complex spectra in comparison to the common Monte Carlo Method joint with the Powell refinement. (author)