返回

定义7设A , B都是n阶矩阵 , 若有可逆矩阵P , 使P-1AP=B

则称B是A的相似矩阵 , 或说矩阵A与B相似

对A进行P-1AP运算称为对A进行相似变换

可逆矩阵P称为把A变成B的相似变换矩阵

举例说明

A=(2103)A = \begin{pmatrix} 2 & 1 \\ 0 & 3 \end{pmatrix}

取一个可逆矩阵 PP =(1101)= \begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix},那么P1=(1101)P^{- 1} = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}

计算P1AP=(1101)(2103)(1101)=(2003)=BP^{- 1}AP = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}\begin{pmatrix} 2 & 1 \\ 0 & 3 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} = \begin{pmatrix} 2 & 0 \\ 0 & 3 \end{pmatrix} = B

所以 AABB 相似,

运算过程就是相似变换。

定理3若n阶矩阵A与B相似 , 则A与B的特征多项式相同

从而A与B的特征值亦相同

证明:

AABB 相似,则存在可逆矩阵 PP,使得B=P1AP.B = P^{- 1}AP.

考虑 BB 的特征多项式:

det(BλI)=det(P1APλI)=det(P1APλP1P)=det(P1(AλI)P).det(B - \lambda I) = det(P^{- 1}AP - \lambda I) = det(P^{- 1}AP - \lambda P^{- 1}P) = det(P^{- 1}(A - \lambda I)P).

利用行列式性质 det(XY)=det(X)det(Y)det(XY) = det(X)det(Y) 以及 det(P1)=1det(P)det(P^{- 1}) = \frac{1}{det(P)}

det(P1(AλI)P)=det(P1)det(AλI)det(P)det(P^{- 1}(A - \lambda I)P) = det(P^{- 1}) \cdot det(A - \lambda I) \cdot det(P)

=det(P1)det(P)det(AλI)= det(P^{- 1}) \cdot det(P) \cdot det(A - \lambda I)

=1det(AλI).= 1 \bullet det(A - \lambda I).

因此det(BλI)=det(AλI).det(B - \lambda I) = det(A - \lambda I).

AABB 的特征多项式相同,所以它们的特征值也相同。

证毕。

例:二阶矩阵验证

取矩阵A=(2112)A = \begin{pmatrix} 2 & 1 \\ 1 & 2 \end{pmatrix},可逆矩阵P=(1101)P = \begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix},先求P1P^{- 1}

二阶矩阵逆公式:若P=(abcd)P = \begin{pmatrix} a & b \\ c & d \end{pmatrix},则P1=1adbc(dbca)P^{- 1} = \frac{1}{ad - bc}\begin{pmatrix} d & - b \\ - c & a \end{pmatrix},得P1=(1101)P^{- 1} = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}

步骤1:求相似矩阵B=P1APB = P^{- 1}AP

P1AP=(1101)(2112)(1101)=(1112)(1101)=(1013)=BP^{- 1}AP = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}\begin{pmatrix} 2 & 1 \\ 1 & 2 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} = \begin{pmatrix} 1 & - 1 \\ 1 & 2 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} = \begin{pmatrix} 1 & 0 \\ 1 & 3 \end{pmatrix} = B

步骤2:求AA的特征多项式与特征值

特征多项式:|λEA|=|λ211λ2|=(λ2)21=λ24λ+3|\lambda E - A| = \left| \begin{matrix} \lambda - 2 & - 1 \\ - 1 & \lambda - 2 \end{matrix} \right| = (\lambda - 2)^{2} - 1 = \lambda^{2} - 4\lambda + 3

特征值:令λ24λ+3=0\lambda^{2} - 4\lambda + 3 = 0,解得λ1=1,λ2=3\lambda_{1} = 1,\lambda_{2} = 3

步骤3:求BB的特征多项式与特征值

特征多项式:|λEB|=|λ101λ3|=(λ1)(λ3)=λ24λ+3|\lambda E - B| = \left| \begin{matrix} \lambda - 1 & 0 \\ - 1 & \lambda - 3 \end{matrix} \right| = (\lambda - 1)(\lambda - 3) = \lambda^{2} - 4\lambda + 3

特征值:令(λ1)(λ3)=0(\lambda - 1)(\lambda - 3) = 0,解得λ1=1,λ2=3\lambda_{1} = 1,\lambda_{2} = 3

结论:AABB的特征多项式均为λ24λ+3\lambda^{2} - 4\lambda + 3,特征值均为1133,符合定理。

推论 若n阶矩阵A与对角矩阵Λ=[λ1λ2λn]\left\lbrack \begin{array}{r} \begin{matrix} \lambda_{1} \\ \end{matrix} \\ \begin{matrix} \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \lambda_{2} \end{matrix} \\ \begin{matrix} \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \end{matrix} \\ \begin{matrix} \ddots \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \end{matrix} \\ \begin{matrix} \\ \lambda_{n} \end{matrix} \end{array} \right\rbrack相似

则λ1 , λ2 , ⋯ , λn即是A的n个特征值

证: 若n阶矩阵A与Λ相似 , 则A与Λ的特征多项式相同

从而A与Λ的特征值亦相同,

因为λ1 , λ2 , ⋯ , λn即是Λ的n个特征值,

所以λ1 , λ2 , ⋯ , λn也就是A的n个特征值

例:2阶矩阵验证(单特征值)

步骤1:构造矩阵并求其特征值

取2阶实对称矩阵(必可对角化):A=(2112)A = \begin{pmatrix} 2 & 1 \\ 1 & 2 \end{pmatrix}

特征方程为|λEA|=0|\lambda E - A| = 0,展开计算:

|λEA|=λ211λ2=(λ2)21=λ24λ+3=(λ1)(λ3)=0|\lambda E - A| = \begin{vmatrix} \lambda - 2 & - 1 \\ - 1 & \lambda - 2 \end{vmatrix} = (\lambda - 2)^{2} - 1 = \lambda^{2} - 4\lambda + 3 = (\lambda - 1)(\lambda - 3) = 0

解得AA的特征值:λ1=1\lambda_{1} = 1λ2=3\lambda_{2} = 3

步骤2:验证AA与对角矩阵Λ\Lambda相似

(1)求特征向量构造相似变换矩阵PP

λ1=1\lambda_{1} = 1,解(EA)x=0(E - A)x = 0,得基础解系ξ1=(11)\xi_{1} = \begin{pmatrix} 1 \\ - 1 \end{pmatrix}

λ2=3\lambda_{2} = 3,解(3EA)x=0(3E - A)x = 0,得基础解系ξ2=(11)\xi_{2} = \begin{pmatrix} 1 \\ 1 \end{pmatrix}

构造可逆矩阵P=(ξ1,ξ2)=(1111)P = (\xi_{1},\xi_{2}) = \begin{pmatrix} 1 & 1 \\ - 1 & 1 \end{pmatrix}

其逆矩阵为:P1=1|P|P*=12(1111)P^{- 1} = \frac{1}{|P|}P^{\ast} = \frac{1}{2}\begin{pmatrix} 1 & - 1 \\ 1 & 1 \end{pmatrix}

(2)验证P1AP=ΛP^{- 1}AP = \Lambda

P1AP=12(1111)(2112)(1111)=(1003)=ΛP^{- 1}AP = \frac{1}{2}\begin{pmatrix} 1 & - 1 \\ 1 & 1 \end{pmatrix}\begin{pmatrix} 2 & 1 \\ 1 & 2 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ - 1 & 1 \end{pmatrix} = \begin{pmatrix} 1 & 0 \\ 0 & 3 \end{pmatrix} = \Lambda

结论

AA与对角矩阵Λ=(1003)\Lambda = \begin{pmatrix} 1 & 0 \\ 0 & 3 \end{pmatrix}相似,Λ\Lambda的主对角线元素131、3,恰好是AA的全部特征值,

在第2章中我们曾指出:若A=PBP-1 , 则Ak=PBkP-1 , A的多项式φ(A)=Pφ(B)P-1

特别 , 若有可逆矩阵P使P-1AP=Λ为对角矩阵 , 即若A相似于对角矩阵Λ

则Ak=PΛkP-1 , φ(A)=Pφ(Λ)P-1

而对于对角矩阵Λ , 有Λk=[λ1kλ2kλnk]\left\lbrack \begin{array}{r} \begin{matrix} \lambda_{1}^{k} \\ \end{matrix} \\ \begin{matrix} \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \lambda_{2}^{k} \end{matrix} \\ \begin{matrix} \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \end{matrix} \\ \begin{matrix} \ddots \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \end{matrix} \\ \begin{matrix} \\ \lambda_{n}^{k} \end{matrix} \end{array} \right\rbrack , φ(Λ)=[φ(λ1)φ(λ2)φ(λn)]\left\lbrack \begin{array}{r} \begin{matrix} \varphi\left( \lambda_{1} \right) \\ \end{matrix} \\ \begin{matrix} \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \varphi\left( \lambda_{2} \right) \end{matrix} \\ \begin{matrix} \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \end{matrix} \\ \begin{matrix} \ddots \\ \end{matrix} \end{array}\ \ \ \ \begin{array}{r} \begin{matrix} \\ \end{matrix} \\ \begin{matrix} \\ \varphi\left( \lambda_{n} \right) \end{matrix} \end{array} \right\rbrack

由此可方便地计算A的多项式φ(A)

对n阶矩阵A , 寻求相似变换矩阵P , 使P-1AP=Λ为对角矩阵

这就称为把矩阵A相似对角化

定理4

nn阶矩阵A与对角矩阵相似(即A可相似对角化)A\text{与对角矩阵相似(即}A\text{可相似对角化)}的充分必要条件

An个线性无关的特征向量A\text{有}n\text{个线性无关的特征向量}

例子:验证必要性

A=(2103)A = \begin{pmatrix} 2 & 1 \\ 0 & 3 \end{pmatrix}

第一步:验证 AA 可对角化

取矩阵P=(1101),P1=(1101)P = \begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix},则P^{- 1} = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}

计算P1AP=(1101)(2103)(1101)=(2003)P^{- 1}AP = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}\begin{pmatrix} 2 & 1 \\ 0 & 3 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} = \begin{pmatrix} 2 & 0 \\ 0 & 3 \end{pmatrix}

因此 AA 相似于对角矩阵 B=(2003)B = \begin{pmatrix} 2 & 0 \\ 0 & 3 \end{pmatrix},即 AA 可对角化。

第二步:求 AA 的特征值与特征向量

特征多项式为|AλE|=|2λ103λ|=(2λ)(3λ)=0|A - \lambda E| = \left| \begin{matrix} 2 - \lambda & 1 \\ 0 & 3 - \lambda \end{matrix} \right| = (2 - \lambda)(3 - \lambda) = 0

解得特征值 λ1=2\lambda_{1} = 2λ2=3\lambda_{2} = 3

对于 λ1=2\lambda_{1} = 2: 有方程组 (A2E)x=(0101)[x1x2]=0(A - 2E)x = \begin{pmatrix} 0 & 1 \\ 0 & 1 \end{pmatrix}\left\lbrack \begin{array}{r} x_{1} \\ x_{2} \end{array} \right\rbrack = 0

通解为x=c(10),cx = c\begin{pmatrix} 1 \\ 0 \end{pmatrix},\quad c \in \mathbb{R},取 v1=(10)v_{1} = \begin{pmatrix} 1 \\ 0 \end{pmatrix}

对于 λ2=3\lambda_{2} = 3: 有方程组 (A3E)x=(1100)[x1x2]=0(A - 3E)x = \begin{pmatrix} - 1 & 1 \\ 0 & 0 \end{pmatrix}\left\lbrack \begin{array}{r} x_{1} \\ x_{2} \end{array} \right\rbrack = 0

通解为x=c(11),cx = c\begin{pmatrix} 1 \\ 1 \end{pmatrix},\quad c \in \mathbb{R}v2=(11)v_{2} = \begin{pmatrix} 1 \\ 1 \end{pmatrix}

第三步:检验特征向量的线性无关性

v1=(10)v_{1} = \begin{pmatrix} 1 \\ 0 \end{pmatrix}v2=(11)v_{2} = \begin{pmatrix} 1 \\ 1 \end{pmatrix},显然它们不成比例,故线性无关。

第四步:结论

由于 AA 可对角化,我们求出了它确实有两个线性无关的特征向量 v1,v2v_{1},v_{2}

这验证了定理的必要性: “若 AA 可对角化,则 AA 必有 nn 个线性无关的特征向量。”

例子:验证充分性

A=(2103)A = \begin{pmatrix} 2 & 1 \\ 0 & 3 \end{pmatrix}

第一步:求 AA 的特征值与特征向量

特征多项式为|AλE|=|2λ103λ|=(2λ)(3λ)=0|A - \lambda E| = \left| \begin{matrix} 2 - \lambda & 1 \\ 0 & 3 - \lambda \end{matrix} \right| = (2 - \lambda)(3 - \lambda) = 0

解得特征值 λ1=2\lambda_{1} = 2λ2=3\lambda_{2} = 3

对于 λ1=2\lambda_{1} = 2: 有方程组 (A2E)x=(0101)[x1x2]=0(A - 2E)x = \begin{pmatrix} 0 & 1 \\ 0 & 1 \end{pmatrix}\left\lbrack \begin{array}{r} x_{1} \\ x_{2} \end{array} \right\rbrack = 0

通解为x=c(10),cx = c\begin{pmatrix} 1 \\ 0 \end{pmatrix},\quad c \in \mathbb{R},取 v1=(10)v_{1} = \begin{pmatrix} 1 \\ 0 \end{pmatrix}

对于 λ2=3\lambda_{2} = 3: 有方程组 (A3E)x=(1100)[x1x2]=0(A - 3E)x = \begin{pmatrix} - 1 & 1 \\ 0 & 0 \end{pmatrix}\left\lbrack \begin{array}{r} x_{1} \\ x_{2} \end{array} \right\rbrack = 0

通解为x=c(11),cx = c\begin{pmatrix} 1 \\ 1 \end{pmatrix},\quad c \in \mathbb{R}v2=(11)v_{2} = \begin{pmatrix} 1 \\ 1 \end{pmatrix}

第二步:检验特征向量的线性无关性

v1=(10)v_{1} = \begin{pmatrix} 1 \\ 0 \end{pmatrix}v2=(11)v_{2} = \begin{pmatrix} 1 \\ 1 \end{pmatrix},显然它们不成比例,故线性无关。

因此,AA22 个线性无关的特征向量。

第三步:构造可逆矩阵 PP 并对角化

P=(v1,v2)=(1101)P = (v_{1},v_{2}) = \begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix},则 PP 可逆,且P1=(1101)P^{- 1} = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}

P1(AP)=(1101)(2103)(1101)=(2003)P^{- 1}(AP) = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}\begin{pmatrix} 2 & 1 \\ 0 & 3 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} = \begin{pmatrix} 2 & 0 \\ 0 & 3 \end{pmatrix}

第四步:结论

由于我们构造出了可逆矩阵 PP,使得 P1APP^{- 1}AP 为对角矩阵,因此 AA 可相似对角化。

这就验证了定理的充分性:

AAnn 个线性无关的特征向量,则 AA 可对角化。

推论nn 阶矩阵 AAnn 个特征值互不相等,

就一定有n个线性无关的特征向量

从而一定能相似对角化。

举例验证(2阶矩阵)

A=(1102)A = \begin{pmatrix} 1 & 1 \\ 0 & 2 \end{pmatrix}

1. 求特征值

|λIA|=|λ110λ2|=(λ1)(λ2)=0|\lambda I - A| = \left| \begin{matrix} \lambda - 1 & - 1 \\ 0 & \lambda - 2 \end{matrix} \right| = (\lambda - 1)(\lambda - 2) = 0

特征值:λ1=1,λ2=2\lambda_{1} = 1,\ \lambda_{2} = 2,两个不同特征值。

2. 求特征向量

λ1=1\lambda_{1} = 1(IA)ξ=0(0101)(x1x2)=0(I - A)\xi = 0 \Rightarrow \begin{pmatrix} 0 & - 1 \\ 0 & - 1 \end{pmatrix}\begin{pmatrix} x_{1} \\ x_{2} \end{pmatrix} = 0

取一个特征向量:ξ1=(10)\xi_{1} = \begin{pmatrix} 1 \\ 0 \end{pmatrix}

λ2=2\lambda_{2} = 2(2IA)ξ=0(1100)(x1x2)=0(2I - A)\xi = 0 \Rightarrow \begin{pmatrix} 1 & - 1 \\ 0 & 0 \end{pmatrix}\begin{pmatrix} x_{1} \\ x_{2} \end{pmatrix} = 0

取一个特征向量:ξ2=(11)\xi_{2} = \begin{pmatrix} 1 \\ 1 \end{pmatrix}

3. ξ1=(10),ξ2=(11)\xi_{1} = \begin{pmatrix} 1 \\ 0 \end{pmatrix},与\xi_{2} = \begin{pmatrix} 1 \\ 1 \end{pmatrix}线性无关,正好 2 个线性无关特征向量。

4. 可相似对角化

P=(ξ1,ξ2)=(1101)P = (\xi_{1},\xi_{2}) = \begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix},则P1AP=(1101)(1102)(1101)=(1002)P^{- 1}AP = \begin{pmatrix} 1 & - 1 \\ 0 & 1 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ 0 & 2 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} = \begin{pmatrix} 1 & 0 \\ 0 & 2 \end{pmatrix}

确实相似对角化了。

当A的特征方程有重根时 , 就不一定有n个线性无关的特征向量

从而不一定能相似对角化。

例如矩阵A=[110430102]\begin{bmatrix} - 1 & 1 & 0 \\ - 4 & 3 & 0 \\ 1 & 0 & 2 \end{bmatrix}

特征方程|Aλ|=|1λ1043λ0102λ|=(2λ)(1λ)2|A - \lambda| = \left| \begin{matrix} - 1 - \lambda & 1 & 0 \\ - 4 & 3 - \lambda & 0 \\ 1 & 0 & 2 - \lambda \end{matrix} \right| = (2 - \lambda)(1 - \lambda)^{2}

特征值为λ1=2λ2=λ3=1\lambda_{1} = 2,\lambda_{2} = \lambda_{3} = 1

有重根,找不到3个线性无关的特征向量

因此A不能相似对角化

当A的特征方程有重根时 , 也可能有n个线性无关的特征向量,

从而能相似对角化。

例10设矩阵A=[211020413]\begin{bmatrix} - 2 & 1 & 1 \\ 0 & 2 & 0 \\ - 4 & 1 & 3 \end{bmatrix} , 那么A能否相似对角化?

若能 , 则求可逆矩阵P和对角矩阵Λ , 使P-1AP=Λ

解:先求A的特征值

|AλE|=|2λ1102λ0413λ|=(2λ)|2λ143λ||A - \lambda E| = \left| \begin{matrix} - 2 - \lambda & 1 & 1 \\ 0 & 2 - \lambda & 0 \\ - 4 & 1 & 3 - \lambda \end{matrix} \right| = (2 - \lambda)\left| \begin{matrix} - 2 - \lambda & 1 \\ - 4 & 3 - \lambda \end{matrix} \right|

=(2λ)(2 - \lambda)[ λ² - λ - 2 ]=-(λ+1)(λ - 2)2

解得特征值λ1 = -1,λ2 = λ3= 2(二重特征值),

当λ1 = -1时,解方程(A-(-1)E)x=0

由A-(-1)E=[111030414][101010000]\begin{bmatrix} - 1 & 1 & 1 \\ 0 & 3 & 0 \\ - 4 & 1 & 4 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 0 & - 1 \\ 0 & 1 & 0 \\ 0 & 0 & 0 \end{bmatrix}

得同解方程组:{x1x3=0x2=0\begin{matrix} \left\{ \begin{matrix} x_{1} - x_{3} = 0 \\ \ \ \ \ \ \ \ \ \ x_{2} = 0\ \ \ \end{matrix} \right.\ \end{matrix}

得参数形式:{x1=x3x2=0\left\{ \begin{array}{r} x_{1} = x_{3} \\ x_{2} = 0 \end{array} \right.\ (x3 可任意取值)

得参数形式:{x1=cx2=0x3=c\left\{ \begin{matrix} x_{1} = c \\ x_{2} = 0 \\ x_{3} = c \end{matrix} \right.\ (其中x3=c)

得齐次通解:[x1x2x3]\begin{bmatrix} x_{1} \\ x_{2} \\ x_{3} \end{bmatrix}=[c0c]\begin{bmatrix} c \\ 0 \\ c \end{bmatrix}=c[101]c\begin{bmatrix} 1 \\ 0 \\ 1 \end{bmatrix} (c为任意实数)

得与λ=-1对应的特征向量ξ1=[101]\begin{bmatrix} 1 \\ 0 \\ 1 \end{bmatrix}

当λ2 = λ3= 2时,解方程(A-2E)x=0

由A-2E=[411000411][411000000]\begin{bmatrix} - 4 & 1 & 1 \\ 0 & 0 & 0 \\ - 4 & 1 & 1 \end{bmatrix} \rightarrow \begin{bmatrix} - 4 & 1 & 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{bmatrix}

得同解方程组:4x₁−x₂−x₃=0

得参数形式: x1=14x2+14x3x_{1} = \frac{1}{4}x_{2} + \frac{1}{4}x_{3} (x2 , x3可任意取值)

得参数形式:{x1=14c1+14c2x2=c1x3=c2\left\{ \begin{matrix} x_{1} = \frac{1}{4}c_{1} + \frac{1}{4}c_{2} \\ x_{2} = c_{1} \\ x_{3} = c_{2} \end{matrix} \right.\ (其中x2=c1 , x3=c2 )

得齐次通解:[x1x2x3]\begin{bmatrix} x_{1} \\ x_{2} \\ x_{3} \end{bmatrix}=[14c1+14c2c1c2]\begin{bmatrix} \frac{1}{4}c_{1} + \frac{1}{4}c_{2} \\ c_{1} \\ c_{2} \end{bmatrix}=c1[1410]c_{1}\begin{bmatrix} \frac{1}{4} \\ 1 \\ 0 \end{bmatrix}+c2[1401]c_{2}\begin{bmatrix} \frac{1}{4} \\ 0 \\ 1 \end{bmatrix} (c1 , c2为任意实数)

得与λ=2对应的线性无关特征向量为ξ2=[1410]=[140]\begin{bmatrix} \frac{1}{4} \\ 1 \\ 0 \end{bmatrix} = \begin{bmatrix} 1 \\ 4 \\ 0 \end{bmatrix} , ξ3=[1401]=[104]\begin{bmatrix} \frac{1}{4} \\ 0 \\ 1 \end{bmatrix} = \begin{bmatrix} 1 \\ 0 \\ 4 \end{bmatrix}

矩阵A是3阶矩阵,特征向量的并集是线性无关的,

这里我们求得3个线性无关的特征向量ξ₁, ξ₂, ξ₃,所以矩阵A可以相似对角化。

可逆矩阵P由特征向量作为列向量构成,

即P=(ξ₁, ξ₂, ξ₃)=[101140104]\left\lbrack \begin{matrix} 1 \\ 0 \\ 1 \end{matrix}\ \ \ \ \begin{matrix} 1 \\ 4 \\ 0 \end{matrix}\ \ \ \ \begin{matrix} 1 \\ 0 \\ 4 \end{matrix}\ \right\rbrack P-1=[43013131411213013]\left\lbrack \begin{matrix} \frac{4}{3} \\ 0 \\ \frac{- 1}{3} \end{matrix}\ \ \ \ \begin{matrix} \frac{- 1}{3} \\ \frac{1}{4} \\ \frac{1}{12} \end{matrix}\ \ \ \ \begin{matrix} \frac{- 1}{3} \\ 0 \\ \frac{1}{3} \end{matrix}\ \right\rbrack

Λ=P-1AP=[43013131411213013][211020413][101140104]=[100020002]\left\lbrack \begin{matrix} \frac{4}{3} \\ 0 \\ \frac{- 1}{3} \end{matrix}\ \ \ \ \begin{matrix} \frac{- 1}{3} \\ \frac{1}{4} \\ \frac{1}{12} \end{matrix}\ \ \ \ \begin{matrix} \frac{- 1}{3} \\ 0 \\ \frac{1}{3} \end{matrix}\ \right\rbrack\begin{bmatrix} - 2 & 1 & 1 \\ 0 & 2 & 0 \\ - 4 & 1 & 3 \end{bmatrix}\left\lbrack \begin{matrix} 1 \\ 0 \\ 1 \end{matrix}\ \ \ \ \begin{matrix} 1 \\ 4 \\ 0 \end{matrix}\ \ \ \ \begin{matrix} 1 \\ 0 \\ 4 \end{matrix}\ \right\rbrack = \begin{bmatrix} - 1 & 0 & 0 \\ 0 & 2 & 0 \\ 0 & 0 & 2 \end{bmatrix}

上例说明:

当A的特征方程有重根时 ,也可能有n个线性无关的特征向量

也就可相似对角化。

例11设A=[00111t100]\begin{bmatrix} 0 & 0 & 1 \\ 1 & 1 & t \\ 1 & 0 & 0 \end{bmatrix} , 问t为何值时 , 矩阵A能相似对角化?

解:

依据n阶矩阵A可相似对角化的充要条件:A有n个线性无关的特征向量,

对于3阶矩阵,A可相似对角化的充要条件是:A有3个线性无关的特征向量。

步骤1:求矩阵A的所有特征值

矩阵的特征方程为|λEA|=0|\lambda E - A| = 0

其中特征矩阵λEA\lambda E - A为:λEA=(λ011λ1t10λ)\lambda E - A = \begin{pmatrix} \lambda & 0 & - 1 \\ - 1 & \lambda - 1 & - t \\ - 1 & 0 & \lambda \end{pmatrix}

按第二列展开计算行列式(第二列仅有λ1\lambda - 1非零,计算最简):

|λEA|=(λ1)(1)2+2|λ11λ|==(λ1)(λ21)==(λ1)2(λ+1)|\lambda E - A| = (\lambda - 1) \cdot ( - 1)^{2 + 2} \cdot \left| \begin{matrix} \lambda & - 1 \\ - 1 & \lambda \end{matrix} \right| = = (\lambda - 1)(\lambda^{2} - 1) = = (\lambda - 1)^{2}(\lambda + 1)

解得A的特征值为:λ=1\lambda = 1(二重根),λ=1\lambda = - 1(单根)。

步骤2:分析特征向量的数量要求

对应某个特征值的线性无关的特征向量,

就是齐次线性方程组(λEA)x=0(\lambda E - A)x = 0的基础解系中的向量。

1. 对于单根特征值λ=1\lambda = - 1

代入得齐次方程组(EA)x=0( - E - A)x = 0,其系数矩阵为:EA=(10112t101)- E - A = \begin{pmatrix} - 1 & 0 & - 1 \\ - 1 & - 2 & - t \\ - 1 & 0 & - 1 \end{pmatrix}

初等行变换后可得该矩阵的秩为2。

根据齐次线性方程组的性质:

基础解系所含线性无关解的个数 = 未知数个数 - 系数矩阵的秩,

此处未知数个数为3,因此基础解系含32=13 - 2 = 1个线性无关的解,

λ=1\lambda = - 1对应1个线性无关的特征向量,该结果与参数t无关。

2. 对于二重根特征值λ=1\lambda = 1

要让A总共有3个线性无关的特征向量,

必须让λ=1\lambda = 1对应2个线性无关的特征向量(1+2=3,满足3阶矩阵的要求)。

结合齐次线性方程组的性质,此处未知数个数为3,

要让基础解系含2个线性无关的解,必须满足:3r(EA)=23 - r(E - A) = 2

即核心要求为r(EA)=1r(E - A) = 1

步骤3:化简EAE - A,求解参数t

代入λ=1\lambda = 1,得:EA=(10110t101)E - A = \begin{pmatrix} 1 & 0 & - 1 \\ - 1 & 0 & - t \\ - 1 & 0 & 1 \end{pmatrix}

对矩阵做初等行变换(初等行变换不改变矩阵的秩):

变换后的矩阵为:(10100t1000)\begin{pmatrix} 1 & 0 & - 1 \\ 0 & 0 & - t - 1 \\ 0 & 0 & 0 \end{pmatrix}

要让该矩阵的秩为1,必须让矩阵中不全为零的行仅有1行,

因此第2行必须为全零行,即:t1=0- t - 1 = 0

解得t=1t = - 1

结论验证 当t=1t = - 1时,EAE - A初等行变换后为(101000000)\begin{pmatrix} 1 & 0 & - 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{pmatrix},秩为1,

此时(EA)x=0(E - A)x = 0的基础解系含2个线性无关的解,

λ=1\lambda = 1对应2个线性无关的特征向量。

加上λ=1\lambda = - 1对应的1个线性无关的特征向量,A总共有3个线性无关的特征向量,完全满足相似对角化的充要条件。

最终结论 当且仅当t=1t = - 1时,矩阵A能相似对角化。

t=1t = - 1代入验证

例11设A=[001111100]\begin{bmatrix} 0 & 0 & 1 \\ 1 & 1 & - 1 \\ 1 & 0 & 0 \end{bmatrix} , 那么A能否相似对角化?

若能 , 则求可逆矩阵P和对角矩阵Λ , 使P-1AP=Λ

解:先求A的特征值

|AλE|=|0λ0111λ1100λ||A - \lambda E| = \left| \begin{matrix} 0 - \lambda & 0 & 1 \\ 1 & 1 - \lambda & - 1 \\ 1 & 0 & 0 - \lambda \end{matrix} \right|

解得特征值λ1 = -1,λ2 = λ3= 1(二重特征值),

当λ1 = -1时,解方程(A-(-1)E)x=0

由A-(-1)E=[101121101][101011000]\begin{bmatrix} 1 & 0 & 1 \\ 1 & 2 & - 1 \\ 1 & 0 & 1 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 0 & 1 \\ 0 & 1 & - 1 \\ 0 & 0 & 0 \end{bmatrix}

得同解方程组:{x1+x3=0x2x3=0\begin{matrix} \left\{ \begin{matrix} x_{1} + x_{3} = 0 \\ \ \ \ \ \ \ \ \ \ x_{2} - x_{3} = 0\ \ \ \end{matrix} \right.\ \end{matrix}

得参数形式:{x1=x3x2=x3\left\{ \begin{array}{r} x_{1} = - x_{3} \\ x_{2} = x_{3} \end{array} \right.\ (x3 可任意取值)

得参数形式:{x1=cx2=cx3=c\left\{ \begin{matrix} x_{1} = - c \\ x_{2} = c \\ x_{3} = c \end{matrix} \right.\ (其中x3=c)

得齐次通解:[x1x2x3]\begin{bmatrix} x_{1} \\ x_{2} \\ x_{3} \end{bmatrix}=[ccc]\begin{bmatrix} - c \\ c \\ c \end{bmatrix}=c[111]c\begin{bmatrix} - 1 \\ 1 \\ 1 \end{bmatrix} (c为任意实数)

得对应λ=-1的特征向量ξ1=[111]\begin{bmatrix} - 1 \\ 1 \\ 1 \end{bmatrix}

当λ2 = λ3= 1时,解方程(A-E)x=0

由A-E=[101101101][101101000]\begin{bmatrix} - 1 & 0 & 1 \\ 1 & 0 & - 1 \\ 1 & 0 & - 1 \end{bmatrix} \rightarrow \begin{bmatrix} - 1 & 0 & 1 \\ 1 & 0 & - 1 \\ 0 & 0 & 0 \end{bmatrix} [101101000][101000000]\rightarrow \begin{bmatrix} 1 & 0 & - 1 \\ - 1 & 0 & 1 \\ 0 & 0 & 0 \end{bmatrix} \rightarrow \begin{bmatrix} 1 & 0 & - 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \end{bmatrix}

得同解方程组:x₁−x₃=0

得参数形式: x1=x3x_{1} = x_{3} ( x3可任意取值)

得参数形式:{x1=c2x2=c1x3=c2\left\{ \begin{matrix} x_{1} = c_{2} \\ x_{2} = c_{1} \\ x_{3} = c_{2} \end{matrix} \right.\ (其中x2=c1, x3=c2)

得齐次通解:[x1x2x3]\begin{bmatrix} x_{1} \\ x_{2} \\ x_{3} \end{bmatrix}=[c2c1c2]\begin{bmatrix} c_{2} \\ c_{1} \\ c_{2} \end{bmatrix}=c1[010]+c2[101]c_{1}\begin{bmatrix} 0 \\ 1 \\ 0 \end{bmatrix} + c_{2}\begin{bmatrix} 1 \\ 0 \\ 1 \end{bmatrix} (c1,c2为任意实数)

得对应λ=1的线性无关特征向量为ξ2=[010]\begin{bmatrix} 0 \\ 1 \\ 0 \end{bmatrix} , ξ3=[101]\begin{bmatrix} 1 \\ 0 \\ 1 \end{bmatrix}

矩阵A是3阶矩阵,特征向量的并集是线性无关的,

这里我们求得3个线性无关的特征向量ξ₁, ξ₂, ξ₃,所以矩阵A可以相似对角化。

可逆矩阵P由特征向量作为列向量构成,

即P=(ξ₁, ξ₂, ξ₃)=[111010101]\left\lbrack \begin{matrix} {- 1} \\ 1 \\ 1 \end{matrix}\ \ \ \ \begin{matrix} 0 \\ 1 \\ 0 \end{matrix}\ \ \ \ \begin{matrix} 1 \\ 0 \\ 1 \end{matrix}\ \right\rbrack P-1=[101121101]\begin{bmatrix} - 1 & 0 & 1 \\ 1 & 2 & - 1 \\ 1 & 0 & 1 \end{bmatrix}

Λ=P-1AP=[101121101][001111100][111010101]=[200020002]=[100010001]\begin{bmatrix} - 1 & 0 & 1 \\ 1 & 2 & - 1 \\ 1 & 0 & 1 \end{bmatrix}\begin{bmatrix} 0 & 0 & 1 \\ 1 & 1 & - 1 \\ 1 & 0 & 0 \end{bmatrix}\left\lbrack \begin{matrix} {- 1} \\ 1 \\ 1 \end{matrix}\ \ \ \ \begin{matrix} 0 \\ 1 \\ 0 \end{matrix}\ \ \ \ \begin{matrix} 1 \\ 0 \\ 1 \end{matrix} \right\rbrack = \begin{bmatrix} - 2 & 0 & 0 \\ 0 & 2 & 0 \\ 0 & 0 & 2 \end{bmatrix} = \begin{bmatrix} - 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1 \end{bmatrix}

验证了当t=-1时,矩阵A可以对解化。