The least squares problem has many applications in the field of optimization. In the present work, we use two strategies for its resolution: Levenberg-Marquardt and Conjugate Gradients. Each one exploits some problem features, and both are globalized by the trust-region strategy. Our contribution consists in the implementation of both methods using the CAS Maxima and in the comparative analysis of these methods in the resolution of a family of least squares problems from the literature.
Least squares; trust region; computational implementation
