Parallel inexact Newton and interior point method

Publication date 25 February 2002

Published in Annali dell'Università di Ferrara. Nuova Serie. Sezione VII (Search for Journal in Brave)

inexact Newton method linear least squares nonlinear mixed complementary problems numerical tests parallel numerical methods quasi-Newton methods sparse and large systems of nonlinear equations systems of nonlinear equations

Mathematics Subject Classification ID

Numerical computation of solutions to systems of equations (65H10) Numerical mathematical programming methods (65K05) Complementarity and equilibrium problems and variational inequalities (finite dimensions) (aspects of mathematical programming) (90C33) Interior-point methods (90C51)

Review text

An inexact Newton method is combined with a block iterative row-projection linear solver in order to solve sparse and large systems of nonlinear equations. It is underlined that a mutually orthogonal row-partition of the Jacobian matrix allows a simple solution of the linear least squares sub-problems. Numerical tests concerning the usage of the new method for nonlinear problems arising from a discretization of partial differential equations are done on a parallel computer. Interior point methods are used for the solution of nonlinear mixed complementary problems, in a sequential manner.

Recommendations

Cited in

(3)

This page was built for publication: Parallel inexact Newton and interior point method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2718096)