In recent years several proposals for the step-size selection have largely improved the gradient methods, in the case of both constrained and unconstrained nonlinear optimization. We introduce a new step-size rule with some crucial properties. We design step-size selection strategies where the new rule and a standard Barzilai-Borwein (BB) rule can be adaptively alternated to get meaningful convergence rate improvements in comparison with other BB-like gradient schemes.
On adaptive step-size selections in gradient methods
ZANGHIRATI, Gaetano
2007
Abstract
In recent years several proposals for the step-size selection have largely improved the gradient methods, in the case of both constrained and unconstrained nonlinear optimization. We introduce a new step-size rule with some crucial properties. We design step-size selection strategies where the new rule and a standard Barzilai-Borwein (BB) rule can be adaptively alternated to get meaningful convergence rate improvements in comparison with other BB-like gradient schemes.File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.