A posteriori error estimation of goal-oriented quantities for elliptic type BVPs

作者:

Highlights:

摘要

In practice the process or object under analysis is usually modelled by means of a selected mathematical model, whose approximate solution is computed with a help of a certain computer code. This approximate solution necessarily includes various errors related to the approximation itself, special features of the particular method used, round-off errors, etc. Therefore, it inevitably rises the question about the reliability of the computed approximations. In the present paper we describe and test numerically the new effective computational technology designed for a control of the accuracy of approximate solutions in terms of goal-oriented quantities (or goal-oriented criteria). Such quantities are to be chosen by a user depending on solution properties that present a special interest. The technology proposed is applicable to the elliptic type boundary-value problems and leads to effective computer codes aimed to control errors of approximate solutions obtained by the finite element method which presents nowadays the main computational tool in industrial software. Various numerical tests confirming high effectivity of this technology are presented.

论文关键词:65N15,65N30,65N50,A posteriori error estimation,Goal-oriented quantity,Finite element method,Differential equation of elliptic type,Superconvergence

论文评审过程:Received 16 February 2005, Available online 25 October 2005.

论文官网地址:https://doi.org/10.1016/j.cam.2005.06.038