Incomplete LU factorization: Difference between revisions

Content deleted Content added
Bender the Bot (talk | contribs)
m Stability: HTTP → HTTPS for the American Mathematical Society, replaced: =http://www.ams.org/ → =https://www.ams.org/
Adding local short description: "Concept in numerical linear algebra", overriding Wikidata description "sparse approximation of the LU factorization often used as a preconditioner"
 
(2 intermediate revisions by 2 users not shown)
Line 1:
{{Short description|Concept in numerical linear algebra}}
In [[numerical linear algebra]], an '''incomplete LU factorization''' (abbreviated as '''ILU''') of a [[matrix (mathematics)|matrix]] is a [[sparse matrix|sparse]] approximation of the [[LU factorization]] often used as a [[preconditioner]].
 
Line 15 ⟶ 16:
G(A) := \left\lbrace (i,j) \in \N^2 : A_{ij} \neq 0 \right\rbrace \,,
</math>
which is used to define the conditions a ''sparsity patternspattern'' <math> S </math> needs to fulfill
:<math>
S \subset \left\lbrace 1, \dots , n \right\rbrace^2
Line 51 ⟶ 52:
More accurate ILU preconditioners require more memory, to such an extent that eventually the running time of the algorithm increases even though the total number of iterations decreases. Consequently, there is a cost/accuracy trade-off that users must evaluate, typically on a case-by-case basis depending on the family of linear systems to be solved.
 
TheAn approximation to the ILU factorization can be performed as a [[fixed-point iteration]] in a highly parallel way.<ref>{{cite journal|last1=Chow|first1=Edmond|last2=Patel|first2=Aftab|title=Fine-grained parallel incomplete LU factorization|journal=SIAM Journal on Scientific Computing|date=2015|volume=37|issue=2|page=C169-C193|ref=iterativeILU|doi=10.1137/140968896}}</ref>
 
== See also ==