矩阵A可以LU分解的充要条件
矩阵A可以LU分解的充要条件是
∀ k ∈ { 1 , ⋯ , n } , r a n k A 1 : k , 1 : k + k ⩾ r a n k A 1 : k , 1 : n + r a n k A 1 : n , 1 : k \forall k \in \{1,\cdots,n\}, \mathrm{rank} A_{1:k,1:k} + k \geqslant \mathrm{rank} A_{1:k,1:n} + \mathrm{rank} A_{1:n,1:k} ∀k∈{1,⋯,n},rankA1:k,1:k+k⩾rankA1:k,1:n+rankA1:n,1:k.如果矩阵A可以LU分解, 分解的唯一性需要通过下式分析
L y = f Ly = f Ly=f, U x = y Ux = y Ux=y.下面给出一个分解算法.
- INPUT A ∈ M n × n ( F ) A \in \mathcal{M}_{n \times n}(\mathcal{F}) A∈Mn×n(F)
- FOR k = 1 , ⋯ , k k = 1, \cdots, k k=1,⋯,k DO
- IF A = 0 A = 0 A=0 THEN
- L : , k ← 0 L_{:,k} \gets 0 L:,k←0
- U k , : ← 0 U_{k,:} \gets 0 Uk,:←0
- ELSE
- FIND i , j , i,j, i,j, WHICH MINIMIZE min { i , j } \min\{i,j\} min{i,j} SUBJECT TO a i j ≠ 0 a_{ij} \neq 0 aij=0
- L : , k ← a i j − 1 A : , j L_{:,k} \gets a_{ij}^{-1} A_{:,j} L:,k←aij−1A:,j
- U k , : ← A i , : U_{k,:} \gets A_{i,:} Uk,:←Ai,:
- A ← A − L : , k U k , : A \gets A - L_{:,k} U_{k,:} A←A−L:,kUk,:
- END IF
- END FOR
- OUTPUT L L L, U U U,
参考文献
Pavel Okunev, Necessary And Sufficient Conditions For Existence of the LU Factorization of an Arbitrary Matrix.
书上的定理7.3是
矩阵A可以LU分解的充分条件是
A
A
A的顺序主子式
D
i
≠
0
D_i \neq 0
Di=0(
i
=
1
,
2
,
⋯
,
n
−
1
i = 1,2, \cdots, n-1
i=1,2,⋯,n−1)
此时分解是唯一的.