In this paper, the optimization of geodetic observation weights through an eigenvalue problem and two methods for its solution are presented: the classic Newton method (linear estimate) and the quase-Newton - BFGS method which uses quadratic models to minimize a function. It’s presented the basis for the optimization of the geodetic weighted observations formulated through the analysis of the eigenvalues in the covariance matrix estimated of the parameters. The proposed methods are applied in three local bi-dimensional geodetic networks and the results are analysed and discussed.
Optimization of Geodetic Observations Weights; Eigenvalue Problem; Pre-Analysis; Least Squares