An extrapolation accelerated multiscale Newton-MG method for fourth-order compact discretizations of semilinear Poisson equations

An extrapolation accelerated multiscale Newton-multigrid (EMNMG) method is proposed to solve two-dimensional semilinear Poisson equations. The nine-point fourth-order compact schemes are used to approximate the nonlinear Poisson equations. In order to accelerate Newton-MG method for calculating the...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Computers & mathematics with applications (1987) Ročník 113; s. 189 - 197
Hlavní autoři: Hu, Hongling, Li, Ming, Pan, Kejia, Wu, Pinxia
Médium: Journal Article
Jazyk:angličtina
Vydáno: Oxford Elsevier Ltd 01.05.2022
Elsevier BV
Témata:
Algorithms
Discretization
Extrapolation
Finite difference method
Fourth-order compact scheme
Interpolation
Multigrid algorithm
Nonlinear term
Poisson equation
Polynomials
Extrapolation
Poisson equation
Fourth-order compact scheme
Nonlinear term
Multigrid algorithm
ISSN:0898-1221, 1873-7668
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:An extrapolation accelerated multiscale Newton-multigrid (EMNMG) method is proposed to solve two-dimensional semilinear Poisson equations. The nine-point fourth-order compact schemes are used to approximate the nonlinear Poisson equations. In order to accelerate Newton-MG method for calculating the finite difference (FD) solution on the finest grid, a quite good initial guess is constructed from the fourth-order FD solutions at two coarse levels by using Richardson extrapolation and bi-quartic polynomial interpolation, which greatly reduces the number of Newton iterations required. A completed extrapolation technique is adopted to generate a sixth-order extrapolated solution on entire finest grid cheaply. Numerical results are given to show that our proposed EMNMG algorithm can achieve high accuracy and keep less cost simultaneously.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0898-1221
1873-7668
DOI:10.1016/j.camwa.2022.03.003