第六章:线性方程组的数值解
Doolittle分解
- 矩阵 A = ( 1 2 1 − 2 − 1 − 5 0 3 − 1 ) A = \left(\begin{matrix} 1 & 2 & 1 \\ -2 & -1 & -5 \\ 0 & 3 & -1 \end{matrix}\right) A=⎝⎛1−202−131−5−1⎠⎞,的Doolittle分解式为 ( 1 0 0 − 2 1 0 0 1 1 ) ( 1 2 1 0 3 − 3 0 0 2 ) \left(\begin{matrix} 1 & 0 & 0 \\ -2 & 1 & 0 \\ 0 & 1 & 1 \end{matrix}\right)\left(\begin{matrix} 1 & 2 & 1 \\ 0 & 3 & -3 \\ 0 & 0 & 2 \end{matrix}\right) ⎝⎛1−20011001⎠⎞⎝⎛1002301−32⎠⎞.
![image-20201231141525060](https://typora-1301295568.cos.ap-guangzhou.myqcloud.com/img/image-20201231141525060.png)
类似的:P132页
-
用Doolittle分解求解方程组:
( 2 1 1 1 3 2 1 2 2 ) ( x 1 x 2 x 3 ) = ( 12 20 15 ) \left( \begin{array} { l l l } { 2 } & { 1 } & { 1 } \\ { 1 } & { 3 } & { 2 } \\ { 1 } & { 2 } & { 2 } \end{array} \right) \left( \begin{array} { l } { x _ { 1 } } \\ { x _ { 2 } } \\ { x _ { 3 } } \end{array} \right) = \left( \begin{array} { l } { 12 } \\ { 20 } \\ { 15 } \end{array} \right) ⎝⎛211132122⎠⎞⎝⎛x1x2x3⎠⎞=⎝⎛122015⎠⎞
先LU分解:
然后 L U X = b LUX=b LUX=b,令 U X = Y UX=Y UX=Y,求 Y Y Y, L Y = b LY=b LY=b。 Y = L − 1 b Y=L^{-1}b Y=L−1b,得到如下式子:
Y = ( 12 14 3 5 ) Y=\left(\begin{matrix} 12 \\ 14 \\ \frac{3}{5} \end{matrix}\right) Y=⎝⎛121453⎠⎞, U X = Y UX=Y UX=Y, X = U − 1 Y X=U^{-1}Y X=U−1Y,求出 X X X如下:
柯朗分解
类似于上面的求法,比下面的例题求法简单
![image-20210103112757389](https://typora-1301295568.cos.ap-guangzhou.myqcloud.com/img/image-20210103112757389.png)
![image-20210103112835713](https://typora-1301295568.cos.ap-guangzhou.myqcloud.com/img/image-20210103112835713.png)
雅可比迭代法(Jacobi)
P122
-
设线性代数方程组 A x = b Ax=b Ax=b的系数矩阵为
∣ 1 a a a 1 a a a 1 ∣ \left|\begin{matrix} 1 & a & a \\ a & 1 & a \\ a & a & 1 \end{matrix}\right| ∣∣∣∣∣∣1aaa1aaa1∣∣∣∣∣∣,
其中a为实数,试求能使 Jacobi方法收敛的a的取值范围.
![image-20210105201021316](https://typora-1301295568.cos.ap-guangzhou.myqcloud.com/img/image-20210105201021316.png)
![image-20210105201302484](https://typora-1301295568.cos.ap-guangzhou.myqcloud.com/img/image-20210105201302484.png)
D = ∣ 1 0 0 0 1 0 0 0 1 ∣ D=\left|\begin{matrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1 \end{matrix}\right| D=∣∣∣∣∣∣100010001∣∣∣∣∣∣, B = ∣ 0 − a − a − a 0 − a − a − a 0 ∣ B=\left|\begin{matrix} 0 & -a & -a \\ -a & 0 & -a \\ -a & -a & 0 \end{matrix}\right| B=∣∣∣∣∣∣0−a−a−a0−a−a−a0∣∣∣∣∣∣,
求出 B B B的特征值: ∣ λ E − A ∣ = ∣ λ a a a λ a a a λ ∣ = > ∣ λ + 2 a λ + 2 a λ + 2 a a λ a a a λ ∣ = 0 | \lambda E - A | = \left| \begin{array} { c c c } { \lambda} & { a } & { a } \\ { a } & { \lambda } & { a } \\ { a } & { a } & { \lambda } \end{array} \right| => \left| \begin{array} { c c c } { \lambda+2a} & { \lambda+2a } & { \lambda+2a } \\ { a } & { \lambda } & { a } \\ { a } & { a } & { \lambda } \end{array} \right| = 0 ∣λE−A∣=∣∣∣∣∣∣λaaaλaaaλ∣∣∣∣∣∣=>∣∣∣∣∣∣λ+2aaaλ+2aλaλ+2aaλ∣∣∣∣∣∣=0, ==>
( λ + 2 a ) ∗ ( λ − a ) 2 (\lambda+2a)*(\lambda-a)^2 (λ+2a)∗(λ−a)2
解出 λ 1 = − 2 a , λ 2 , 3 = a \lambda_1=-2a,\lambda_{2,3}=a λ1=−2a,λ2,3=a,
谱半径 ρ = max 1 ⩽ i ⩽ n ∣ λ i ∣ = ∣ 2 a ∣ < 1 \rho=\max _{1 \leqslant i \leqslant n}\left|\lambda_{i}\right|=|2a|<1 ρ=max1⩽i⩽n∣λi∣=∣2a∣<1,解出 ∣ a ∣ < 1 / 2 |a|<1/2 ∣a∣<1/2.
- 设有系数矩阵
A = [ 1 2 − 2 1 1 1 2 2 1 ] A=\left[\begin{array}{lll} 1 & 2 & -2 \\ 1 & 1 & 1 \\ 2 & 2 & 1 \end{array}\right] A=⎣⎡112212−211⎦⎤
证明: 对系数矩阵 A , A, A, 雅可比迭代收敛, 而 Gauss-Seidel 迭代不收敛。
由系数矩阵 A A A,可得雅可比迭代矩阵是 B 1 = I − D − 1 A = [ 0 − 2 2 − 1 0 − 1 − 2 − 2 0 ] B_1=I-D^{-1}A=\left[\begin{array}{lll} 0 & -2 & 2 \\ -1 & 0 & -1 \\ -2 & -2 & 0 \end{array}\right] B1=I−D−1A=⎣⎡0−1−2−20−22−10⎦⎤.
求出特征值: λ 1 = λ 2 = λ 3 = 0 \lambda_1=\lambda_{2}=\lambda_3=0 λ1=λ2=λ3=0,谱半径 ρ = max 1 ⩽ i ⩽ n ∣ λ i ∣ = 0 < 1 \rho=\max _{1 \leqslant i \leqslant n}\left|\lambda_{i}\right|=0<1 ρ=max1⩽i⩽n∣λi∣=0<1,收敛。
另外一个判断方法,严格对角占优这样判断。 P128
高斯–赛尔德迭代矩阵 p128
由系数矩阵 A A A,可得高斯—赛尔德迭代矩阵是 B 2 = − ( D + L ) − 1 U = [ 0 − 2 2 0 2 − 3 0 0 2 ] B_2=-(D+L)^{-1}U=\left[\begin{array}{lll} 0 & -2 & 2 \\ 0 & 2 & -3 \\ 0 & 0 & 2 \end{array}\right] B2=−(D+L)−1U=⎣⎡000−2202−32⎦⎤.
求出特征值: λ 1 = 0 \lambda_1=0 λ1=0, λ 2 , 3 = 2 \lambda _ { 2,3 } = 2 λ2,3=2,谱半径 ρ = max 1 ⩽ i ⩽ n ∣ λ i ∣ = 2 > 1 \rho=\max _{1 \leqslant i \leqslant n}\left|\lambda_{i}\right|=2>1 ρ=max1⩽i⩽n∣λi∣=2>1,不收敛。