线性代数复习笔记
∣
A
∣
=
∑
j
=
1
n
a
i
j
A
i
j
(
i
=
1
,
2
,
.
.
.
,
n
)
=
∑
i
=
1
n
a
i
j
A
i
j
(
j
=
1
,
2
,
.
.
.
,
n
)
;
A
i
j
=
(
−
1
)
i
+
j
M
i
j
.
|A|=\sum_{j=1}^na_{ij}A_{ij}\ (i=1,2,...,n)=\sum_{i=1}^na_{ij}A_{ij}\ (j=1,2,...,n);\ A_ij=(-1)^{i+j}M_{ij}.
∣A∣=∑j=1naijAij (i=1,2,...,n)=∑i=1naijAij (j=1,2,...,n); Aij=(−1)i+jMij.
∑
j
=
1
n
a
i
j
A
k
j
=
∑
i
=
1
n
a
i
j
A
i
k
=
0
,
i
≠
k
.
\sum_{j=1}^na_{ij}A_{kj}=\sum_{i=1}^na_{ij}A_{ik}=0,\ i\ne k.
∑j=1naijAkj=∑i=1naijAik=0, i=k.
副对角线:
∣
B
n
∣
=
(
−
1
)
n
(
n
−
1
)
2
∏
i
=
1
n
b
i
,
n
+
1
−
i
|B_n|=(-1)^\frac{n(n-1)}{2}\prod_{i=1}^nb_{i,n+1-i}
∣Bn∣=(−1)2n(n−1)∏i=1nbi,n+1−i.
范德蒙:
∣
D
n
∣
=
∏
1
≤
j
≤
i
≤
n
(
x
i
−
x
j
)
|D_n|=\prod_{1\leq j\leq i\leq n}(x_i-x_j)
∣Dn∣=∏1≤j≤i≤n(xi−xj).
矩阵:
n
×
n
n\times n
n×n 方阵构成有零因子非交换环; 单位元
∣
E
∣
=
1
|E|=1
∣E∣=1; 零因子
∣
Z
∣
=
0
|Z|=0
∣Z∣=0.
反对称:
a
i
j
+
a
j
i
=
0
⟹
a
i
i
=
0
a_{ij}+a_{ji}=0\implies a_{ii}=0
aij+aji=0⟹aii=0.
转置:
a
i
j
T
=
a
j
i
a^T_{ij}=a_{ji}
aijT=aji;
(
A
+
B
)
T
=
A
T
+
B
T
(A+B)^T=A^T+B^T
(A+B)T=AT+BT;
(
A
B
)
T
=
B
T
A
T
(AB)^T=B^TA^T
(AB)T=BTAT.
逆元:
A
A
−
1
=
A
−
1
A
=
E
⟺
∣
A
∣
≠
0
AA^{-1}=A^{-1}A=E\iff |A|\ne 0
AA−1=A−1A=E⟺∣A∣=0;
∣
A
−
1
∣
=
∣
A
∣
−
1
|A^{-1}|=|A|^{-1}
∣A−1∣=∣A∣−1;
A
−
1
=
∣
A
∣
−
1
A
∗
A^{-1}=|A|^{-1}A^*
A−1=∣A∣−1A∗;
(
A
B
)
−
1
=
B
−
1
A
−
1
(AB)^{-1}=B^{-1}A^{-1}
(AB)−1=B−1A−1.
[
A
∣
E
]
→
∏
i
=
1
k
Q
k
[
E
∣
A
−
1
]
[A|E]\xrightarrow[]{\prod_{i=1}^kQ_k}[E|A^{-1}]
[A∣E]∏i=1kQk[E∣A−1].
伴随:
A
A
∗
=
A
∗
A
=
∣
A
∣
E
AA^*=A^*A=|A|E
AA∗=A∗A=∣A∣E;
∣
A
∗
∣
=
∣
A
∣
n
−
1
|A^*|=|A|^{n-1}
∣A∗∣=∣A∣n−1;
A
∗
=
∣
A
∣
A
−
1
A^*=|A|A^{-1}
A∗=∣A∣A−1;
(
A
∗
)
∗
=
∣
A
∣
n
−
2
A
(
n
≥
2
)
(A^*)^*=|A|^{n-2}A\ (n\geq 2)
(A∗)∗=∣A∣n−2A (n≥2).
正交:
A
A
T
=
A
T
A
=
E
⟹
∣
A
∣
=
±
1
AA^T=A^TA=E\implies |A|=\pm 1
AAT=ATA=E⟹∣A∣=±1, 行(列)向量均为单位向量并两两正交.
初等变换: 对换(
E
i
j
E_{ij}
Eij); 倍乘(
E
i
(
k
)
,
k
≠
0
E_i(k),\ k\ne 0
Ei(k), k=0); 倍加(
E
i
j
(
K
)
E_{ij}(K)
Eij(K)); 左乘为行变换, 右乘为列变换.
∣
E
i
j
A
∣
=
∣
A
E
i
j
∣
=
−
∣
A
∣
|E_{ij}A|=|AE_{ij}|=-|A|
∣EijA∣=∣AEij∣=−∣A∣;
∣
E
i
(
k
)
A
∣
=
∣
A
E
i
(
k
)
∣
=
k
∣
A
∣
|E_i(k)A|=|AE_i(k)|=k|A|
∣Ei(k)A∣=∣AEi(k)∣=k∣A∣;
∣
k
A
∣
=
k
n
∣
A
∣
|kA|=k^n|A|
∣kA∣=kn∣A∣;
∣
E
i
j
(
k
)
A
∣
=
∣
A
E
i
j
(
k
)
∣
=
∣
A
∣
|E_{ij}(k)A|=|AE_{ij}(k)|=|A|
∣Eij(k)A∣=∣AEij(k)∣=∣A∣.
等价:
A
≅
B
⟺
∃
∣
P
∣
,
∣
Q
∣
≠
0
A\cong B\iff\exists |P|,|Q|\ne 0
A≅B⟺∃∣P∣,∣Q∣=0 s.t.
P
A
Q
=
B
PAQ=B
PAQ=B.
秩: 非零子式最大阶数; 极大线性无关组中向量个数.
r
(
A
)
=
0
⟺
A
=
O
r(A)=0\iff A=O
r(A)=0⟺A=O.
r
(
A
)
=
1
⟺
A
≠
O
r(A)=1\iff A\ne O
r(A)=1⟺A=O 且任意两行(列)成比例.
r
(
A
A
T
)
=
r
(
A
T
A
)
=
r
(
A
)
r(AA^T)=r(A^TA)=r(A)
r(AAT)=r(ATA)=r(A).
r
(
A
n
)
<
n
⟺
∣
A
n
∣
=
0
⟺
r(A_n)<n\iff|A_n|=0\iff
r(An)<n⟺∣An∣=0⟺ 不可逆
⟺
\iff
⟺ 有零特征值.
r
(
A
n
)
=
n
⟺
∣
A
n
∣
≠
0
⟺
r(A_n)=n\iff|A_n|\ne 0\iff
r(An)=n⟺∣An∣=0⟺ 可逆
⟺
\iff
⟺ 无零特征值.
r
(
A
±
B
)
≤
r
(
A
)
+
r
(
B
)
r(A\pm B)\leq r(A)+r(B)
r(A±B)≤r(A)+r(B);
r
(
A
B
)
≤
min
{
r
(
A
)
,
r
(
B
)
}
r(AB)\leq\min\{r(A),r(B)\}
r(AB)≤min{r(A),r(B)};
max
{
r
(
A
)
,
r
(
B
)
}
≤
r
(
A
∣
B
)
≤
r
(
A
)
+
r
(
B
)
\max\{r(A),r(B)\}\leq r(A|B)\leq r(A)+r(B)
max{r(A),r(B)}≤r(A∣B)≤r(A)+r(B).
r
(
A
m
×
n
)
=
n
⟹
r
(
A
B
)
=
r
(
B
)
r(A_{m\times n})=n\implies r(AB)=r(B)
r(Am×n)=n⟹r(AB)=r(B).
r
(
C
n
×
s
)
=
n
⟹
r
(
B
C
)
=
r
(
B
)
r(C_{n\times s})=n\implies r(BC)=r(B)
r(Cn×s)=n⟹r(BC)=r(B).
A
≅
B
⟹
r
(
A
)
=
r
(
B
)
A\cong B\implies r(A)=r(B)
A≅B⟹r(A)=r(B).
A
m
×
n
B
n
×
s
=
O
⟹
r
(
A
)
+
r
(
B
)
≤
n
A_{m\times n}B_{n\times s}=O\implies r(A)+r(B)\leq n
Am×nBn×s=O⟹r(A)+r(B)≤n.
r
(
A
∗
)
=
{
n
,
r
(
A
)
=
n
1
,
r
(
A
)
=
n
−
1
0
,
r
(
A
)
<
n
−
1
r(A^*)=\begin{cases}n, & r(A)=n \\ 1, & r(A)=n-1 \\ 0, & r(A)<n-1\end{cases}
r(A∗)=⎩
⎨
⎧n,1,0,r(A)=nr(A)=n−1r(A)<n−1.
α
,
β
≠
0
\bm{\alpha},\bm{\beta}\ne\bm{0}
α,β=0 为等维列向量
⟹
r
(
α
β
T
)
=
1
\implies r(\bm{\alpha}\bm{\beta}^T)=1
⟹r(αβT)=1.
线性表出:
{
α
i
}
i
=
1
m
\{\bm{\alpha}_i\}_{i=1}^m
{αi}i=1m,
β
\bm{\beta}
β,
∃
{
k
i
}
i
=
1
m
\exists \{k_i\}_{i=1}^m
∃{ki}i=1m s.t.
β
=
∑
i
=
1
m
k
i
α
i
⟺
r
(
α
i
)
i
=
1
m
=
r
(
{
α
i
}
i
=
1
m
∣
β
)
⟺
(
α
i
)
i
=
1
m
x
=
β
\bm{\beta}=\sum_{i=1}^m k_i\bm{\alpha}_i\iff r(\bm{\alpha}_i)_{i=1}^m=r(\{\bm{\alpha}_i\}_{i=1}^m|\beta)\iff(\bm{\alpha}_i)_{i=1}^m\bm{x}=\bm{\beta}
β=∑i=1mkiαi⟺r(αi)i=1m=r({αi}i=1m∣β)⟺(αi)i=1mx=β 有解.
r
(
A
)
=
r
(
A
∣
B
)
>
r
(
B
)
⟹
r(A)=r(A|B)>r(B)\implies
r(A)=r(A∣B)>r(B)⟹ 向量组
B
B
B 可由向量组
A
A
A 线性表出, 向量组
A
A
A 不可由向量组
B
B
B 线性表出.
向量组
A
,
B
A,B
A,B 可互相线性表出
⟺
A
≅
B
⟺
r
(
A
)
=
r
(
A
∣
B
)
=
r
(
B
)
\iff A\cong B\iff r(A)=r(A|B)=r(B)
⟺A≅B⟺r(A)=r(A∣B)=r(B).
线性无关:
{
α
i
}
i
=
1
m
\{\bm{\alpha}_i\}_{i=1}^m
{αi}i=1m, 仅当
{
k
i
}
i
=
1
m
=
0
\{k_i\}_{i=1}^m=0
{ki}i=1m=0 时才有
∑
i
=
1
m
k
i
α
i
=
0
⟺
r
(
α
i
)
i
=
1
m
=
m
⟺
(
α
i
)
i
=
1
m
x
=
0
\sum_{i=1}^m k_i\bm{\alpha}_i=\bm{0}\iff r(\bm{\alpha}_i)_{i=1}^m=m\iff (\bm{\alpha}_i)_{i=1}^m\bm{x}=\bm{0}
∑i=1mkiαi=0⟺r(αi)i=1m=m⟺(αi)i=1mx=0 仅有零解.
全体组线性无关
⟹
\implies
⟹ 部分组线性无关.
缩短组线性无关
⟹
\implies
⟹ 延伸组线性无关.
向量组
B
t
B_t
Bt 线性无关, 可由向量组
A
s
A_s
As 线性表出
⟹
s
≥
t
\implies s\geq t
⟹s≥t.
向量组
A
A
A 线性无关, 添加
β
\bm{\beta}
β 后线性相关
⟹
β
\implies\bm{\beta}
⟹β 可由向量组
A
A
A 线性表出且方式唯一.
v
=
A
x
=
B
y
\bm{v}=A\bm{x}=B\bm{y}
v=Ax=By.
过渡矩阵: 基
A
A
A 变换为基
B
B
B, 有
B
=
A
P
B=AP
B=AP.
坐标变换: 坐标
x
\bm{x}
x 变换为
y
\bm{y}
y, 即
y
=
P
−
1
x
\bm{y}=P^{-1}\bm{x}
y=P−1x.
施密特正交化:
β
1
=
α
1
\bm{\beta}_1=\bm{\alpha}_1
β1=α1;
β
i
=
α
i
−
∑
j
=
1
r
−
1
[
α
i
,
β
j
]
[
β
j
,
β
j
]
,
i
=
2
,
3
,
.
.
.
,
r
\bm{\beta}_i=\bm{\alpha}_i-\sum_{j=1}^{r-1}\frac{[\bm{\alpha}_i,\bm{\beta}_j]}{[\bm{\beta}_j,\bm{\beta}_j]},\ i=2,3,...,r
βi=αi−∑j=1r−1[βj,βj][αi,βj], i=2,3,...,r.
单位化:
γ
i
=
β
i
∣
∣
β
i
∣
∣
,
i
=
1
,
2
,
.
.
.
,
r
\bm{\gamma}_i=\frac{\bm{\beta}_i}{||\bm{\beta}_i||},\ i=1,2,...,r
γi=∣∣βi∣∣βi, i=1,2,...,r.
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 仅有零解(唯一解)
⟺
r
(
A
)
=
n
\iff r(A)=n
⟺r(A)=n.
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 有非零解(无穷解)
⟺
r
(
A
)
<
n
\iff r(A)<n
⟺r(A)<n.
A
x
=
b
A\bm{x}=\bm{b}
Ax=b 有唯一解
⟺
r
(
A
)
=
r
(
A
∣
b
)
=
n
\iff r(A)=r(A|\bm{b})=n
⟺r(A)=r(A∣b)=n.
A
x
=
b
A\bm{x}=\bm{b}
Ax=b 有无穷解
⟺
r
(
A
)
=
r
(
A
∣
b
)
<
n
\iff r(A)=r(A|\bm{b})<n
⟺r(A)=r(A∣b)<n.
A
x
=
b
A\bm{x}=\bm{b}
Ax=b 无解
⟺
r
(
A
)
<
r
(
A
∣
b
)
\iff r(A)<r(A|\bm{b})
⟺r(A)<r(A∣b), 即
r
(
A
)
=
r
(
A
∣
b
)
−
1
r(A)=r(A|\bm{b})-1
r(A)=r(A∣b)−1.
基础解系:
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 的
s
=
n
−
r
(
A
)
s=n-r(A)
s=n−r(A) 个线性无关解
{
ξ
i
}
i
=
1
s
\{\bm{\xi}_i\}_{i=1}^s
{ξi}i=1s; 通解
x
=
∑
i
=
1
s
k
i
ξ
i
\bm{x}=\sum_{i=1}^sk_i\bm{\xi}_i
x=∑i=1skiξi.
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 的解的线性组合也为
A
x
=
0
A\bm{x}=0
Ax=0 的解.
ξ
\bm{\xi}
ξ 为
A
x
=
0
A\bm{x}=0
Ax=0 的解,
η
{\eta}
η 为
A
x
=
b
A\bm{x}=\bm{b}
Ax=b 的解
⟹
η
+
k
ξ
\implies\bm{\eta}+k\bm{\xi}
⟹η+kξ 为
A
x
=
b
A\bm{x}=\bm{b}
Ax=b 的解.
A
x
=
b
A\bm{x}=\bm{b}
Ax=b 的解的系数和为
0
0
0 的线性组合为
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 的解; 系数和为
1
1
1 的线性组合为
A
x
=
b
A\bm{x}=\bm{b}
Ax=b 的解.
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 和
B
x
=
0
B\bm{x}=\bm{0}
Bx=0 有非零公共解
⟹
r
(
A
B
)
<
n
\implies r{A\choose B}<n
⟹r(BA)<n.
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 的解均为
B
x
=
0
B\bm{x}=\bm{0}
Bx=0 的解
⟹
r
(
A
)
≥
r
(
B
)
\implies r(A)\geq r(B)
⟹r(A)≥r(B).
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 和
B
x
=
0
B\bm{x}=\bm{0}
Bx=0 同解
⟹
r
(
A
)
=
r
(
B
)
=
r
(
A
B
)
\implies r(A)=r(B)=r{A\choose B}
⟹r(A)=r(B)=r(BA).
A
T
A
x
=
0
A^TA\bm{x}=\bm{0}
ATAx=0 和
A
x
=
0
A\bm{x}=\bm{0}
Ax=0 同解.
A
B
x
=
0
AB\bm{x}=\bm{0}
ABx=0 和
B
x
=
0
B\bm{x}=\bm{0}
Bx=0 同解;
A
A
A 为列满秩.
A
n
n
x
=
0
A_n^n\bm{x}=\bm{0}
Annx=0 和
A
n
n
+
1
x
=
0
A_n^{n+1}\bm{x}=\bm{0}
Ann+1x=0 同解.
特征值:
A
α
=
λ
α
A\bm{\alpha}=\lambda\bm{\alpha}
Aα=λα;
∣
A
−
λ
E
∣
=
0
|A-\lambda E|=0
∣A−λE∣=0.
正交矩阵特征值为
±
1
\pm 1
±1.
f
(
A
)
=
0
⟹
f
(
λ
)
=
0
f(A)=0\implies f(\lambda)=0
f(A)=0⟹f(λ)=0.
特征向量:
(
A
−
λ
E
)
x
=
0
(A-\lambda E)\bm{x}=0
(A−λE)x=0.
k
k
k 重特征值至多有
k
k
k 个线性无关的特征向量; 不同特征值对应的特征向量线性无关; 一个特征向量对应一个特征值; 全部线性无关的特征向量的线性组合也为该特征值的特征向量.
矩阵 | 特征值 | 特征向量 |
---|---|---|
A A A | λ \lambda λ | α \bm{\alpha} α |
A + k E A+kE A+kE | λ + k \lambda+k λ+k | α \bm{\alpha} α |
k A kA kA | k λ k\lambda kλ | α \bm{\alpha} α |
A k A^k Ak | λ k \lambda^k λk | α \bm{\alpha} α |
f ( A ) f(A) f(A) | f ( λ ) f(\lambda) f(λ) | α \bm{\alpha} α |
A − 1 A^{-1} A−1 | 1 λ \frac{1}{\lambda} λ1 | α \bm{\alpha} α |
A ∗ A^* A∗ | ∣ A ∣ λ \frac{|A|}{\lambda} λ∣A∣ | α \bm{\alpha} α |
A T A^T AT | λ \lambda λ | - |
P − 1 A P P^{-1}AP P−1AP | λ \lambda λ | P − 1 α P^{-1}\bm{\alpha} P−1α |
相似:
A
∼
B
⟺
∃
∣
P
∣
≠
0
A\sim B\iff\exists|P|\ne 0
A∼B⟺∃∣P∣=0 s.t.
P
−
1
A
P
=
B
P^{-1}AP=B
P−1AP=B.
A
∼
B
⟹
∣
A
−
λ
E
∣
=
∣
B
−
λ
E
∣
⟹
A
≅
B
A\sim B\implies |A-\lambda E|=|B-\lambda E|\implies A\cong B
A∼B⟹∣A−λE∣=∣B−λE∣⟹A≅B,
t
r
(
A
)
=
t
r
(
B
)
{\rm tr}(A)={\rm tr}(B)
tr(A)=tr(B),
∣
A
∣
=
∣
B
∣
|A|=|B|
∣A∣=∣B∣, 各阶主子式之和分别相等.
A
n
∼
Λ
⟺
A
A_n\sim\Lambda\iff A
An∼Λ⟺A 有
n
n
n 个线性无关的特征向量, 此时
P
=
(
α
i
)
i
=
1
n
P=(\bm{\alpha}_i)_{i=1}^n
P=(αi)i=1n,
Λ
=
P
−
1
A
P
⟺
A
\Lambda=P^{-1}AP\iff A
Λ=P−1AP⟺A 的
k
k
k 重特征根恰有
k
k
k 个线性无关的特征向量, 即
r
(
A
−
λ
E
)
=
n
−
k
r(A-\lambda E)=n-k
r(A−λE)=n−k.
A
n
A_n
An 有
n
n
n 个不同的特征值
⟹
A
∼
Λ
\implies A\sim\Lambda
⟹A∼Λ.
(
A
−
λ
1
E
)
(
A
−
λ
2
E
)
=
O
,
λ
1
≠
λ
2
⟹
A
∼
Λ
(A-\lambda_1 E)(A-\lambda_2 E)=O,\ \lambda_1\ne\lambda_2\implies A\sim\Lambda
(A−λ1E)(A−λ2E)=O, λ1=λ2⟹A∼Λ.
A
=
P
−
1
Λ
P
⟹
A
k
=
P
−
1
Λ
k
P
A=P^{-1}\Lambda P\implies A^k=P^{-1}\Lambda^k P
A=P−1ΛP⟹Ak=P−1ΛkP.
A
=
(
λ
E
+
B
)
⟹
A
k
=
∑
i
=
1
k
(
k
i
)
λ
i
B
k
−
i
A=(\lambda E+B)\implies A^k=\sum_{i=1}^k{k\choose i}\lambda^iB^{k-i}
A=(λE+B)⟹Ak=∑i=1k(ik)λiBk−i;
B
j
=
O
B^j=O
Bj=O,
j
≪
k
j\ll k
j≪k.
A
=
α
β
T
⟹
A
k
=
c
k
−
1
A
A=\bm{\alpha}\bm{\beta}^T\implies A^k=c^{k-1}A
A=αβT⟹Ak=ck−1A,
c
=
β
T
α
c=\bm{\beta}^T\bm{\alpha}
c=βTα.
λ
n
=
∣
λ
E
−
A
∣
Q
(
λ
)
+
R
(
λ
)
⟹
A
n
=
R
(
A
)
\lambda^n=|\lambda E-A|Q(\lambda)+R(\lambda)\implies A^n=R(A)
λn=∣λE−A∣Q(λ)+R(λ)⟹An=R(A).
A
A
A 为实对称矩阵:
a
i
j
=
a
j
i
∈
R
a_{ij}=a_{ji}\in\mathbb{R}
aij=aji∈R;
特征值均为实数;
不同特征值的特征向量正交;
∃
\exists
∃ 正交矩阵
Q
Q
Q s.t.
Q
−
1
A
Q
=
Q
T
A
Q
=
Λ
Q^{-1}AQ=Q^TAQ=\Lambda
Q−1AQ=QTAQ=Λ, 即可正交相似对角化;
B
B
B 也为实对称矩阵,
A
∼
B
⟹
{
λ
A
}
=
{
λ
B
}
A\sim B\implies \{\lambda_A\}=\{\lambda_B\}
A∼B⟹{λA}={λB}.
二次型:
∑
i
=
1
n
∑
j
=
1
n
a
i
j
x
i
x
j
=
x
T
A
x
\sum_{i=1}^n\sum_{j=1}^na_{ij}x_ix_j=\bm{x}^TA\bm{x}
∑i=1n∑j=1naijxixj=xTAx;
A
A
A 为实对称矩阵.
标准形: 只含平方项.
规范形: 只含平方项, 系数只能为
0
,
1
,
−
1
0,1,-1
0,1,−1.
正交变换:
x
=
Q
y
\bm{x}=Q\bm{y}
x=Qy, 即
Q
T
A
Q
=
Λ
Q^TAQ=\Lambda
QTAQ=Λ; 正交矩阵
Q
=
(
γ
i
=
1
n
)
Q=(\bm{\gamma}_{i=1}^n)
Q=(γi=1n) 为
A
A
A 特征向量的正交单位化.
配方换元:
x
=
C
y
\bm{x}=C\bm{y}
x=Cy,
C
C
C 为可逆矩阵.
惯性定理: 标准形正负系数个数(正负惯性指数)分别等于正负特征值个数, 数量和为矩阵秩.
合同:
A
≃
B
⟺
∃
∣
C
∣
≠
0
A\simeq B\iff\exists |C|\ne 0
A≃B⟺∃∣C∣=0 s.t.
B
=
C
T
A
C
⟺
A
,
B
B=C^TAC\iff A,B
B=CTAC⟺A,B 的正负惯性指数分别相等.
正定:
∀
x
≠
O
\forall \bm{x}\ne O
∀x=O s.t.
x
T
A
x
>
0
\bm{x}^TA\bm{x}>0
xTAx>0.
A
n
A_n
An 正定
⟺
\iff
⟺ 正惯性指数为
n
⟺
n\iff
n⟺ 标准形平方项系数全为正
⟺
\iff
⟺ 特征值均大于
0
⟺
0\iff
0⟺ 各阶顺序主子式均大于
0
⟺
∃
∣
P
∣
≠
0
0\iff\exists |P|\ne 0
0⟺∃∣P∣=0 s.t.
A
=
P
T
P
A=P^TP
A=PTP, 即
A
≃
E
⟹
a
i
i
>
0
,
i
=
1
,
2
,
.
.
.
,
n
A\simeq E\implies a_ii>0,\ i=1,2,...,n
A≃E⟹aii>0, i=1,2,...,n,
∣
A
∣
>
0
|A|>0
∣A∣>0.