Inexact Restoration method for nonlinear optimization without derivatives

MARÍA B. AROUXÉT; NÉLIDA ECHEBEST; ELVIO A. PILOTTA

Córdoba

Congreso; IV Congreso Latinoamericano de Matemáticos; 2012

UMALCA - FaMAF-Universidad Nacional de Córdoba

Numerical methods for optimization without derivatives are needed when objective function and the constraints are not available to be used. We propose a new method for nonlinear programming problems without derivatives based on the Inexact Restoration method, which was introduced in [1, 2]. These methods proceed in two phases and gener-ate a sequence of generally infeasible iterates with intermediate iterations that consist of inexactly restored points. In our approach, all the derivatives calculations are performed by using polynomial interpolation and trust-region strategies. We also present numerical experiments.