A local-relaxation method for solving convection-diffusion equations
For constant-coefficient equations Young's SOR theory can be used to find an explicit formula for the optimum relaxation factor, even in cases when diagonal dominance is heavily violated. When desired simple approximations of ωopt can be derived. Combining these with the LR strategy leads to a simple method which for varying coefficient equations, and for nonlinear equations, is able to outperform the optimum SOR method. The LR method can easily solve central-difference approximations of convection-diffusion equations, thus eliminating the need for the trade-off in favour of the upwind-type schemes mentioned in the introduction.
KeywordsSpectral Radius Relaxation Factor Diagonal Dominance Homogeneous Dirichlet Condition National Aerospace Laboratory
Unable to display preview. Download preview PDF.
- D.M. Young, Iterative solution of large linear systems, Academic Press, New York, 1971.Google Scholar
- D. Dijkstra, Ph. D. thesis, University of Groningen, 1974.Google Scholar
- C.W. Richards and C.M. Crane, Appl. Math. Modelling 2 (1978) 59–61.Google Scholar
- D.B. Russell, ARC Report R & M 3331, Oxford, 1963.Google Scholar
- A. Rigal, J. Comput. Phys. 32 (1979) 10–23.Google Scholar
- E.F.F. Botta and A.E.P. Veldman, to be published in J. Comput. Phys.; also NLR MP 81052 U (1981).Google Scholar
- N. Takemitsu, J. Comput. Phys. 36 (1980) 236–248.Google Scholar
- J. Strikwerda, SIAM J. Sci. Stat. Comput. 1 (1980) 119–130.Google Scholar
- L.W. Ehrlich, J. Comput. Phys. 44 (1981) 31–45.Google Scholar
- L.A. Kurtz, R.E. Smith, C.L. Parks and L.R. Boney, Comp. Fluids 6 (1978) 49–70.Google Scholar