Fitting Lorentzian peaks with evolutionary genetic algorithm based on stochastic search procedure

A global search technique for curve fitting based on evolutionary random search was modified and applied for quantifying a combination of Gaussian and Lorentzian peaks. This stochastic search procedures based on randomized operators is a modified Monte Carlo method. The proposed method tested on self obtained several overlapped Lorentzian peaks with random noise, Lennard particles in three dimensions and discrete mathematical functions previously used for optimization in literature. It was found to be the proposed method is suitable for complex and large scale optimization. The results of the new method have been compared with those obtained by two peak fitting programs. Developed method was found to be very fast and thus it is time saving.