The Johnson´s SB probability distribution has wide use in forestry. There are basically five methods for fitting the location (e) and the scale (l) parameters as independent terms that must be known to obtain the other parameters. This work was carried out aiming at developing a new methodology to optimize the determination of the location and scale parameters of this distribution to minimize the "dn" statistics of the adherence test of Kolmogorov-Smirnov by the five fitting methods available. In order to test the performance of the new methodology the "Solver.xla" of the Microsoft Excel 2000 was utilized, defining the objective function and the restrictions of each fitting method. As conclusion of the study, it was noticed that the proposed methodology may be considered as an interesting alternative to the Johnson´s SB function though the optimization schedule. Therefore, the methodology can be widely recommended to obtaining the model parameters of this probability distribution used in forestry.
Johnson´s S B distribution; Fitting methods; Probability distribution and Optimization
