RESTARTED NONLINEAR CONJUGATE GRADIENT METHOD FOR PARA METE IDENTIFICATION IN ELLIPTIC SYSTEM

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

View graph of relations

Author(s)

Detail(s)

Original languageEnglish
Pages (from-to)62-77
Journal / PublicationEurasian Journal of Mathematical and Computer Applications
Volume1
Issue number1
Publication statusPublished - 2013
Externally publishedYes

Abstract

We propose a restarting technique for the nonlinear conjugate gradient (NCG) method in solving the inverse problem of parameter identification in elliptic systems. The technique can enhance the performance of the NCG method in a certain generic situation. We conduct extensive numerical experiments to compare the performance of five well-known conjugate gradient schemes by incorporating the restarting technique. Numerical results show the efficiency and stable performance of the restarted NCG schemes. The smoothing property of the proposed method is also investigated, which could be employed in the multigrid solver for inverse problems.

Research Area(s)

  • Nonlinear conjugate gradient method, inverse problems, parameter identification, finite element methods