--
前面有2篇博文:矩阵迭代, 稀疏矩阵相关。初学的认识。最近又来咀嚼回头草了,仍然参考两本书:
Iterative Methods for linear sparse systems
Lectcutres on computational numerical analysis of PDE. J.M.McDonough, Kentucky.
第一本是hpc课,老师主推荐;第二本的作者,写了几个lectures都非常好,e.g. <cfd>, <turbulenence>
- 经典迭代方法
* Jacobi, Gauss-Seidel, SOR serials :
- to solve Au=b, decompose A = D - L - U
D: dialogue part of A, L : lower part of A; U : upper part of A
u^(n+1) = G u^n + k
- for convergence test, contrast fixed-point iteration form, and through contraction mapping principle( 压缩映射定理). requiring spectral of system less than 1 .
namely rho(G) < 1
Jacobi G_ja = D;
G-S G_gs = D - L;
SOR G_sor = (D- wL)/w, w is relaxation paramter
SSOR G_ssor = (D - wL)D^(-1)(D - wF)/ (w(2-w))
* ADI
decompose A = H + V + S
H: x-direction term; V: y/z-direction term; S: 0-order term