Title :
An Affine Scaling Interior Method for Solving System of Bound-Constrained Semismooth Equations
Author_Institution :
Inst. of Gen. Educ., Sanda Univ., Shanghai, China
Abstract :
In this paper, we present an affine scaling interior algorithm for solving bound-constrained semi-smooth equations. In this algorithm, we develop Inexact Newton Method that is combined with a line search technique. The affine technique and step back-tracking along inexact Newton steps are used. If iteration direction doesn´t satisfy rules expected, the method allows switch to new step in which both line search and interior point backtracking techniques decrease function values. The paper presents a full proof of the method with both global and local super-linear convergences.
Keywords :
Newton method; affine transforms; backtracking; linear algebra; Inexact Newton Method; affine scaling interior method; affine technique; global super-linear convergences; interior point backtracking techniques; iteration direction; line search technique; local super-linear convergences; step back-tracking; system of bound-constrained semismooth equation solving; Convergence; Educational institutions; Equations; Internet; Jacobian matrices; Minimization; Newton method; Semi-smooth; bound constrained; global convergence; interior point; local convergence;
Conference_Titel :
Internet Computing for Engineering and Science (ICICSE), 2013 Seventh International Conference on
Conference_Location :
Shanghai
DOI :
10.1109/ICICSE.2013.21