Numerical experiments with variations of the Gauss-Newton algorithm for nonlinear least squares

Spedicato, E.; Vespucci, M. T.

doi:10.1007/BF00938543

In this paper, the classical Gauss-Newton method for the unconstrained least squares problem is modified by introducing a quasi-Newton approximation to the second-order term of the Hessian. Various quasi-Newton formulas are considered, and numerical experiments show that most of them are more efficient on large residual problems than the Gauss-Newton method and a general purpose minimization algorithm based upon the BFGS formula. A particular quasi-Newton formula is shown numerically to be superior. Further improvements are obtained by using a line search that exploits the special form of the function.

(1988). Numerical experiments with variations of the Gauss-Newton algorithm for nonlinear least squares [journal article - articolo]. In JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS. Retrieved from http://hdl.handle.net/10446/138869

Titolo:	Numerical experiments with variations of the Gauss-Newton algorithm for nonlinear least squares
Tipologia specifica:	articolo
Tutti gli autori:	Spedicato, E.; Vespucci, M. T.
Data di pubblicazione:	1988
Abstract (eng):	In this paper, the classical Gauss-Newton method for the unconstrained least squares problem is modified by introducing a quasi-Newton approximation to the second-order term of the Hessian. Various quasi-Newton formulas are considered, and numerical experiments show that most of them are more efficient on large residual problems than the Gauss-Newton method and a general purpose minimization algorithm based upon the BFGS formula. A particular quasi-Newton formula is shown numerically to be superior. Further improvements are obtained by using a line search that exploits the special form of the function.
Rivista:	JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS
Nelle collezioni:	1.1.01 Articoli/Saggi in rivista - Journal Articles/Essays

File allegato/i alla scheda:

Non ci sono file allegati a questa scheda.

Utilizza questo identificativo per citare o creare un link a questo documento:
http://hdl.handle.net/10446/138869

Citazioni