On the Ka canov Method for a Quasi-Newtonian Flow Problem 1.pdfVIP

On the Kacanov Method for a Quasi-NewtonianFlow Problem 1Jiuhua Chen2, Weimin Han3 and Hong-ci Huang4Abstract. The Kacanov method is an iteration method for solving nonlinearproblems. In this paper, we apply the Kacanov method to solve the problemof quasi-Newtonian ow obeying the Carreau law. The emphasis of thepaper is the derivation of a posteriori error estimate for Kacanov iterates. Anumerical example shows the eciency of the error estimate.1 IntroductionThe Kacanov method is an iteration method for solving nonlinear problems, via lin-earization. Its applications in solving nonlinear problems can be traced back to threedecades ago or earlier. A brief survey on early literature on the Kacanov method canbe found in the introduction section of [8]. Convergence of the method in general formsis given in [10] and [11]. In [7], convergence of the method is proved for solving a familyof even more general problems, which involve nonlinear variational inequalities of themixed kind (i.e., the problems are posed over non-empty, closed, convex subsets and con-tain non-dierentiable terms), and the result is applied in solving a nonlinear variationalinequality of the second kind arising in elastoplasticity. In [8], the Kacanov method isused to solve two model nonlinear boundary value problems, where some a posteriorierror estimates are derived for the Kacanov iterates, and the eciency of the error esti-mates is shown through several numerical examples. In this paper, we extend some ofthe earlier discussions to cover nonlinear boundary value problems with constraints. We1The work was supported by Research Grants Council of Hong Kong UGC.2Department of Mathematics, University of Iowa, Iowa City, IA 52242, U.S.A.3Department of Mathematics, University of Iowa, Iowa City, IA 52242, U.S.A.4Department of Mathematics, Hong Kong Baptist University, Kowloon, Hong Kong.1 will take as a sample problem that of a quasi-Newtonian ow, where the constraint isthe incompressibility