Abstract
A new method for automatic fitting of Pline profiles in UV spectra of stellar winds is presented. The line source function is calculated using Sobolev's approximation and the emergent flux is obtained by exact integration of the equation of the radiation transport (similar to the SEI method described by Lamers et al. (1987)). The quality of the fit is evaluated using the likelihood estimator. The maximization of the likelihood is done by a genetic algorithm. The advantages of our method with respect to other similar approaches are its robustness and its insensibility to the initial guess. In addition, the algorithm guarantees the localization of the global maximum of the likelihood hypersurface, which is not the case for classical minimization algorithms. Here we present an implementation of the genetic algorithm for line profile fitting, its tests on both synthetic and real data and an estimation of the confidence limits of the results.
