线性代数 05.09 相似矩阵及二次型 习题课

相似矩阵及二次型主要知识网络图
$\left \{ \begin{array}{l} 6.1 向量的内积 \\ 6.2 特征值与特征向量 \\ 6.3 二次型 \end{array} \right.$

6.1 向量的内积
$\left \{ \begin{array}{l} 定义: [x, y] = \sum x_iy _i \\ 性质:\left \{ \begin{array}{l} 1.[x, y] = [y, x] \\ 2.\lambda[x, y] = [\lambda x, y] \\ 3.[x + y, z] = [x, z] + [y, z] \end{array} \right. \\ 范数: \Vert x \Vert = \sqrt {[x, x]} \\ 夹角: \theta = \arccos {\dfrac{[x, y]}{\Vert x \Vert \Vert y \Vert } }\\ 正交:[x, y] = 0 \end{array} \right.$

6.2 特征值与特征向量
$\left \{ \begin{array}{l} 定义:Ax = \lambda x, x \neq 0 \\ 求法: \left \{ \begin{array}{l} 特征值: \left \{ \begin{array}{l} 1.定义法; \\ 2.特征多项式法|A - \lambda E|. \end{array} \right. \\ 特征向量: \left \{ \begin{array}{l} 1.定义法; \\ (A - \lambda E) x = 0 的基础解系法. \end{array} \right. \end{array} \right. \\ 性质: \left \{ \begin{array}{l} 不同特征值的特征向量线性无关 \\ k重特征值至多有k个线性无关的特征向量 \\ |A| = \lambda_1 \lambda_2 \cdots \lambda_n,\quad \sum a_{ii} = \sum \lambda_i \end{array} \right. \\ 相似:\left \{ \begin{array}{l} 定义:P^{-1}AP = B \\ 可对角化\left \{ \begin{array}{l} \left \{ \begin{array}{l} 1.A有n个线性无关的特征向量 \\ 2.R(A - \lambda_k E) = n -k, \lambda_k是A的k重特征值. \end{array} \right. \\ \left \{ \begin{array}{l} 1.A有n个不同的特征值 \\ 2.A是实对称矩阵. \end{array} \right. \end{array} \right. \\ 应用\left \{ \begin{array}{l} 1.A^n = P \Lambda^n P^{-1} \\ 2.f = x^TAx \Longrightarrow f = y^T \Lambda y. \end{array} \right. \end{array} \right. \\ 实对称矩阵隐含的信息:\left \{ \begin{array}{l} 必可以对角化,且可用正交变换 \\ 不同特征值所对应的特征向量正交 \\ 特征值全为实数 \\ k重特征值必有k个线性无关的特征向量 \\ 与对角矩阵合同 \end{array} \right. \end{array} \right.$

6.3 二次型
$\left \{ \begin{array}{l} 矩阵表示:f = x^TAx \\ 标准型:\left \{ \begin{array}{l} 定义:f = y^T \Lambda y \\ 化标准型. \left \{ \begin{array}{l} 1.正交化方法 \\ 2.配方法 \\ 3.合同变换法 \end{array} \right.\end{array} \right. \\ 正定二次型:\left \{ \begin{array}{l} 惯性定律:\left \{ \begin{array}{l} 惯性指数R(A) = r, \\ 正惯性指数p; \\ 负惯性指数q \end{array} \right. \\ 定义:\forall x \neq 0, x^TAx > 0 \\ 充要条件:\left \{ \begin{array}{l} 特征值全大于零; \\ 正惯性指数为n; \\ 顺序主子式全大于零; \\ A合同E, 或A = U^TU,其中U可逆. \end{array} \right. \\ 必要条件:|A| > 0 \end{array} \right. \end{array} \right.$

$例1.已知\alpha_1 = \begin{pmatrix} 1 \\ 1 \\ 1 \end{pmatrix},求一组非零向量\alpha_2, \alpha_3,使\alpha_1, \alpha_2, \alpha_3两两正交.$
$解:\because \alpha_2, \alpha_3与\alpha_1正交, \\ \therefore \alpha_1^T \alpha_2 = 0; \alpha_1^T \alpha_3 = 0 \\ 即\alpha_2, \alpha_3为方程\alpha_1^T X = 0的解,其中X = \begin{pmatrix} x_1 \\ x_2 \\ x_3 \end{pmatrix}, \\ 亦即方程 x_1 + x_2 + x_3 = 0 \\ 即 x_1 = -x_2 -x_3 \\ 显然,方程组的基础解系为: \\ \xi_2 = \begin{pmatrix} -1 \\ 1 \\ 0 \end{pmatrix}, \xi_3 = \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix}, \\ 把基础解系正交化: \\ \alpha_2 = \xi_2, \alpha_3 = \xi_3 - \dfrac{[\xi_3, \xi_2] \xi_2}{\Vert \xi_2 \Vert ^2 } \\ \alpha_2 = \begin{pmatrix} -1 \\ 1 \\ 0 \end{pmatrix}, \alpha_3 = \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix} - \dfrac{1}{2}\begin{pmatrix} -1 \\ 1 \\ 0 \end{pmatrix} = \begin{pmatrix} -\dfrac{1}{2} \\ -\dfrac{1}{2} \\ 1 \end{pmatrix} \\ 即为所求.$

$例2.设0是矩阵A = \begin{pmatrix} 1 & 0 & 1 \\ 0 & 2 & 0 \\ 1 & 0 & a \\ \end{pmatrix} 的特征值,求a.再求A的其他特征值.$
$解:由于0是a的特征值,所以|A| = 0, \\ 即|A| = 2(a - 1) = 0,从而a = 1 \\ 通过 |A - \lambda E| = 0求A的所有特征值,由 \\ |A - \lambda E| = \begin{pmatrix} 1 - \lambda & 0 & 1 \\ 0 & 2 - \lambda & 0 \\ 1 & 0 & 1 - \lambda \\ \end{pmatrix} \\ = (2 - \lambda)[( 1- \lambda)^2 - 1] \\ = -\lambda(\lambda - 2)^2 = 0 \\ 故 \lambda_1 = 0, \lambda_2 = \lambda_3 = 2, \\ 所以,A的其它特征值为2.$

$例3.设A = \begin{pmatrix} 1 & 2 & 2 \\ 2 & 1 & 2 \\ 2 & 2 & 1 \\ \end{pmatrix} 求A的全部特征值和对应的特征向量.$
$解: ①根据|A - \lambda E | = 0,求A的全部特征值. \\ |A - \lambda E | = \begin{vmatrix} 1 - \lambda & 2 & 2 \\ 2 & 1 - \lambda & 2 \\ 2 & 2 & 1 - \lambda \\ \end{vmatrix} \\ = \begin{vmatrix} 5 - \lambda & 2 & 2 \\ 5 - \lambda & 1 - \lambda & 2 \\ 5 - \lambda & 2 & 1 - \lambda \\ \end{vmatrix} \\ = (5 - \lambda)\begin{vmatrix} 1 & 2 & 2 \\ 1 & 1 - \lambda & 2 \\ 1 & 2 & 1 - \lambda \\ \end{vmatrix} \\ = (5 - \lambda)\begin{vmatrix} 1 & 2 & 2 \\ 0 & -1 - \lambda & 0 \\ 0 & 0 & -1 - \lambda \\ \end{vmatrix} \\ = (5 - \lambda)(-1 - \lambda)^2 \\ = -(\lambda + 1)^2(\lambda - 5) = 0 \\ 解得A的特征值为:\lambda_1 =\lambda_2 = -1, \lambda_3 = 5. \\ ②通过方程(A - \lambda E) x = 0,求A的特征向量, \\ 当\lambda_1 =\lambda_2 = -1时, (A + E) x = 0, 由 \\ A + E = \begin{pmatrix} 2 & 2 & 2 \\ 2 & 2 & 2 \\ 2 & 2 & 2 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 1 & 1 \\ 0 & 0 & 0 \\ 0 & 0 & 0 \\ \end{pmatrix} \\ 解得(A + E) x = 0基础解系 \xi_1 = \begin{pmatrix} -1 \\ 1 \\ 0 \end{pmatrix}, \xi_2 = \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix}, \\ 故 矩阵A对应于特征值\lambda_1 = \lambda_2 = -1的全部特征向量为:\\ k_1 \xi_1 + k_2 \xi_2, (k_1, k_2 \in R) \\ 当\lambda_3 = 5时, (A - 5E)x = 0,由 \\ (A - 5E) = \begin{pmatrix} -4 & 2 & 2 \\ 2 & -4 & 2 \\ 2 & 2 & -4 \\ \end{pmatrix} \sim \begin{pmatrix} 2 & 2 & -4 \\ 2 & -4 & 2 \\ -4 & 2 & 2 \\ \end{pmatrix} \\ \sim \begin{pmatrix} 1 & 1 & -2 \\ 1 & -2 & 1 \\ -2 & 1 & 1 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 1 & -2 \\ 0 & -3 & 3 \\ 0 & 3 & -3 \\ \end{pmatrix} \\ \sim \begin{pmatrix} 1 & 0 & -1 \\ 0 & 1 & -1 \\ 0 & 0 & 0 \\ \end{pmatrix} \\ 故(A - 5E)x = 0的基础解系为: \xi_3 = \begin{pmatrix}1 \\ 1 \\ 1 \end{pmatrix}, \\ 矩阵A对应于特征值\lambda_3 = 5的全部特征向量为k_3 \xi_3, k_3 \in R$

$例4.设3阶矩阵A的特征值为1, -1, 2, 且B = A^3 - 5A^2,试求:\\ (1) 矩阵B的特征值及其相似对角矩阵,并说明理由; \\ (2) 行列式|B|及|A - 5E|.$
$解: (1) 因 Ax = \lambda x时,有 \\ A^3x = \lambda^3x, A^2x = \lambda^2 x, \\ 从而 Bx = (A^3 - 5A^2)x = \lambda^3x - 5\lambda^2 x = (\lambda^3 - 5\lambda^2)x, \\ 由于A的特征值是1, -1, 2, \\ 从而B的特征值为:-4, -6, -12. \\ (2) 因为B的特征值为:-4, -6, -12,所以\\ |B| = (-4) \times (-6) \times (-12) = - 288 . \\ 有因为: |A| = 1 \times (-1) \times 2 = -2. \\ 而|B| = |A^3 - 5A^2| = |A|^2 |A - 5E| \\ 故 |A - 5E| = -288 \div (-2)^2 = -72.$

$例5.设矩阵A与B相似,其中\\ A = \begin{pmatrix} -2 & 0 & 0 \\ 2 & x & 2 \\ 3 & 1 & 1 \\ \end{pmatrix}, B = \begin{pmatrix} -1 & 0 & 0 \\ 0 & 2 & 0 \\ 3 & 0 & y \\ \end{pmatrix} \\ (1) 求x和y的值; \\ (2) 求可逆矩阵P,使P^{-1}AP = B.$
$解: (1) 因为 A \sim B,故其特征多项式相同,即 \\ |A - \lambda E| = |B - \lambda E| \\ \begin{vmatrix} -2 - \lambda & 0 & 0 \\ 2 & x - \lambda & 2 \\ 3 & 1 & 1 - \lambda \\ \end{vmatrix} = \begin{vmatrix} -1 - \lambda & 0 & 0 \\ 0 & 2 - \lambda & 0 \\ 3 & 0 & y - \lambda \\ \end{vmatrix} \\ (-2 - \lambda)[(x - \lambda)(1 - \lambda) - 2] = (-1 - \lambda)(2 - \lambda)(y - \lambda) \\ (\lambda + 2)[\lambda^2 - (x + 1) \lambda + (x - 2)] = (\lambda + 1)(\lambda - 2)(\lambda - y) \\ 令\lambda = -2, 解得y = -2,\\ 令\lambda = -1, 解得x = 0 \\ 即: x = 0, y = -2 \\ (2) 由(1)知\\ A = \begin{pmatrix} -2 & 0 & 0 \\ 2 & 0 & 2 \\ 3 & 1 & 1 \\ \end{pmatrix}, B = \begin{pmatrix} -1 & 0 & 0 \\ 0 & 2 & 0 \\ 3 & 0 & -2 \\ \end{pmatrix} \\ 由于A \sim B,从而A的特征值为 \\ \lambda_1 = -1, \lambda_2 = 2, \lambda_3 = -2. \\ 当\lambda_1 = -1时, 由(A + E) x = 0得\\ (A + E) = \begin{pmatrix} -1 & 0 & 0 \\ 2 & 1 & 2 \\ 3 & 1 & 2 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 0 & 0 \\ 0 & 1 & 2 \\ 0 & 0 & 0 \\ \end{pmatrix} \\ 对应的特征向量为p_1 = \begin{pmatrix} 0 \\ -2 \\ 1 \end{pmatrix}; \\ 当\lambda_1 = 2时, 由(A -2 E) x = 0得\\ (A - 2 E) = \begin{pmatrix} -4 & 0 & 0 \\ 2 & -2 & 2 \\ 3 & 1 & -1 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 0 & 0 \\ 0 & 1 & -1 \\ 0 & 0 & 0 \\ \end{pmatrix} \\ 对应的特征向量为p_2 = \begin{pmatrix} 0 \\ 1 \\ 1 \end{pmatrix}; \\ 当\lambda_1 = -2时, 由(A + 2 E) x = 0得\\ (A + 2E) = \begin{pmatrix} 0 & 0 & 0 \\ 2 & 2 & 2 \\ 3 & 1 & 3 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 0 & 1 \\ 0 & 1 & 0 \\ 0 & 0 & 0 \\ \end{pmatrix} \\ 对应的特征向量为p_3 = \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix}; \\ 由特征向量拼成可逆矩阵P \\ P = \begin{pmatrix} p_1 & p_2 & p_3 \end{pmatrix} = \begin{pmatrix} 0 & 0 & -1 \\ -2 & 1 & 0 \\ 1 & 1 & 1 \\ \end{pmatrix} \\ 使得P^{-1}AP = B$

$例6.设H = E - 2xx^T, x^Tx = 1,\\ 其中x为n维列向量,试证: \\ (1) H是对称矩阵; \\ (2) H是正交矩阵.$
$证: (1) H^T = (E -2xx^T) = E^T - (2xx^T)^T \\ = E -2(x^T)^Tx^T = E - 2xx^T = H, \\ \therefore H是对称矩阵. \\ (2) H^TH = HH = (E -2xx^T) (E -2xx^T) \\ = E - 2xx^T - 2xx^T + 2xx^T \cdot 2xx^T \\ = E - 4xx^T + 4x(x^Tx)x^T \\ = E - 4xx^T + 4xx^T \\ = E \\ \therefore H是正交矩阵.$

$例7.设A = \begin{pmatrix} 1 & 2 & 3 \\ 2 & 1 & 3 \\ 3 & 3 & 6 \\ \end{pmatrix}, \\ 求正交矩阵P,使得P^{-1}AP成为对角矩阵.$
$解: ①根据|A - \lambda E | = 0,求A的全部特征值: \\ |A - \lambda E | = \begin{vmatrix} 1 - \lambda & 2 & 3 \\ 2 & 1 - \lambda & 3 \\ 3 & 3 & 6 - \lambda \\ \end{vmatrix} \\ = \begin{vmatrix} 6 - \lambda & 2 & 3 \\ 6 - \lambda & 1 - \lambda & 3 \\ 12 - \lambda & 3 & 6 - \lambda \\ \end{vmatrix} \\ = \begin{vmatrix} 6 - \lambda & 2 & 3 \\ 0 & -1 - \lambda & 0 \\ 12 - \lambda & 3 & 6 - \lambda \\ \end{vmatrix} \\ =-(1 + \lambda)[(6 - \lambda)^2 - 3(12 - \lambda)] \\ =(1 + \lambda)\lambda(\lambda- 9) = 0 \\ 解得A的全部特征值为 \lambda_1 = -1, \lambda_2 = 0, \lambda_3 = 9 \\ ②根据(A - \lambda E) x = 0,求A的特征向量. \\ 当\lambda = -1时, (A + E) x = 0, 由 \\ A + E = \begin{pmatrix} 2 & 2 & 3 \\ 2 & 2 & 3 \\ 3 & 3 & 7 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 1 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & 1 \\ \end{pmatrix} \\ 解得基础解系为 p_1 = \begin{pmatrix} -1 \\ 1 \\ 0 \end{pmatrix}, \\ 单位化 e_1 = \dfrac{p_1}{ \Vert p_1 \Vert} \begin{pmatrix} -\dfrac{1}{\sqrt{2}} \\ \dfrac{1}{\sqrt{2}} \\ 0 \end{pmatrix} \\ 当\lambda_2 = 0时, (A)x = 0 , 由 \\ A = \begin{pmatrix} 1 & 2 & 3 \\ 2 & 1 & 3 \\ 3 & 3 & 6 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 2 & 3 \\ 0 & -3 & -3 \\ 0 & -3 & -3 \\ \end{pmatrix} \\ \sim \begin{pmatrix} 1 & 0 & 1 \\ 0 & 1 & 1 \\ 0 & 0 & 0 \\ \end{pmatrix} \\ 解得基础解系 p_2 = \begin{pmatrix} -1 \\ -1 \\ 1 \end{pmatrix}, 单位化 e_2 = \begin{pmatrix} -\dfrac{1}{\sqrt{3}} \\ -\dfrac{1}{\sqrt{3}} \\ \dfrac{1}{\sqrt{3}} \end{pmatrix} \\ 当\lambda_3 = 9时, (A - 9E) x = 0, 由 \\ A - 9E = \begin{pmatrix} -8 & 2 & 3 \\ 2 & -8 & 3 \\ 3 & 3 & -3 \\ \end{pmatrix} \sim \begin{pmatrix} 1 & 0 & -\dfrac{1}{2} \\ 0 & 1 & -\dfrac{1}{2} \\ 0 & 0 & 0 \\ \end{pmatrix} \\ 得基础解系p_3 = \begin{pmatrix} 1 \\ 1\\ 2 \end{pmatrix}, 单位化 e_3 = \begin{pmatrix} \dfrac{1}{\sqrt{6}} \\ \dfrac{1}{\sqrt{6}} \\ \dfrac{2}{\sqrt{6}} \end{pmatrix} \\ 正交矩阵P = \begin{pmatrix} e_1, e_2, e_3 \end{pmatrix} = \begin{pmatrix} -\dfrac{1}{\sqrt{2}} & -\dfrac{1}{\sqrt{3}} & \dfrac{1}{\sqrt{6}} \\ \dfrac{1}{\sqrt{2}} & -\dfrac{1}{\sqrt{3}} & \dfrac{1}{\sqrt{6}} \\ 0 & \dfrac{1}{\sqrt{3}} & \dfrac{2}{\sqrt{6}} \\ \end{pmatrix} \\ 则P^{-1}AP = \begin{pmatrix} -1 & & \\ & 0 & \\ & & 9 \\ \end{pmatrix}$

$例8.设A是n阶正定矩阵,证明:|A + 2E | > 2^n.$
$证1：因为A为正定矩阵,所以存在正交矩阵P,使\\ P^{-1}AP = \Lambda = \begin{pmatrix} \lambda_1 & & & \\ & \lambda_2 & & \\ & & \ddots & \\ & & & \lambda_n \\ \end{pmatrix} \\ 其中\lambda_i为矩阵A的特征值.因为A是正定矩阵, \\ 所以, \lambda_i > 0(i = 1, 2, \cdots, n) \\ 而 P^{-1}(A + 2E)P \\ = P^{-1}AP + P^{-1}(2E)P \\ = \Lambda + 2E \\ = \begin{pmatrix} \lambda_1 + 2 & & & \\ & \lambda_2 + 2 & & \\ & & \ddots & \\ & & & \lambda_n + 2 \\ \end{pmatrix} \\ 故 |A + 2E| = |P^{-1}(A+2E)P| \\ = (\lambda_1+ 2)(\lambda_2 + 2) \cdots (\lambda_n + 2) \\ 因为 \lambda_i > 0, (i = 1, 2, \cdots, n) \\ |A + 2E| > 2^n. \\ 证2： 因为A是正定矩阵,所以\\ A的全部特征值\lambda_i(i = 1, 2, \cdots, n)都大于零. \\ 又因为A + 2E是矩阵A的多项式, \\ 故 A + 2E的特征值\lambda_i + 2(i = 1, 2, \cdots, n)\\ 由矩阵的特征值与其对应的行列式之间的关系,可得\\ |A + 2E | = (\lambda_1 + 2)(\lambda_2 + 2) \cdots (\lambda_n + 2) > 2^n.$

$例9.用不同方法判别A = \begin{pmatrix} 2 & 2 & -2 \\ 2 & 5 & -4 \\ -2 & -4 & 5 \\ \end{pmatrix}的正定性.$
$解1.(惯性指数法) \\ f(x_1, x_2, x_3) = 2x_1^2 + 5x_2^2 + 5x_3^2 + 4x_1x_2 - 4x_1x_3 - 8x_2x_3 \\ = 2(x_1^2 + 2x_1x_2 - 2x_1x_3) + 5x_2^2 + 5x_3^2 - 8 x_2x_3 \\ = 2(x_1 + x_2 - x_3)^2 + 3(x_2 - \dfrac{2}{3}x_3)^2 + \dfrac{23}{9}x_3^2$
$令\left \{ \begin{array}{l} y_1 = x_1 + x_2 - x_3 \\ y_2 = x_2 - \dfrac{2}{3}x_3 \\ y_3 = x_3 \end{array} \right.$
$则得标准型\\ f = 2y_1^2 + 3y_2^2 + \dfrac{23}{9}y_3^2 .\\ 由标准型可知,正惯性指数为3. \\ 所以,f为正定二次型,其矩阵A是正定的.$
$解2:(顺序主子式法) \\ A的各阶顺序主子式为\\ |A_1| = \begin{vmatrix} 2 \\ \end{vmatrix} = 2 > 0; \\ |A_2| = \begin{vmatrix} 2 & 2\\ 2 & 5\\ \end{vmatrix} = 6 > 0 \\ |A_3| = \begin{vmatrix} 2 & 2 & -2 \\ 2 & 5 & -4 \\ -2 & -4 & 5 \\ \end{vmatrix} = \begin{vmatrix} 2 & 2 & -2 \\ 0 & 3 & -2 \\ 0 & -2 & 3 \\ \end{vmatrix} \\ = 2 \times ( 9 - 4) = 10 > 0 \\ 所以，A是正定矩阵. \\ 解3:(特征值法)\\ 由于 \\ |A - \lambda E| = \begin{vmatrix} 2 - \lambda & 2 & -2 \\ 2 & 5 - \lambda & -4 \\ -2 & -4 & 5 - \lambda \\ \end{vmatrix} \\ = \begin{vmatrix} 2 - \lambda & 2 & -2 \\ 0 & 1 - \lambda & 1 - \lambda \\ -2 & -4 & 5 - \lambda \\ \end{vmatrix} \\ = (1 - \lambda) \begin{vmatrix} 2 - \lambda & 2 & -2 \\ 0 & 1 & 1 \\ -2 & -4 & 5 - \lambda \\ \end{vmatrix} \\ = (1 - \lambda) \begin{vmatrix} 2 - \lambda & 2 & -4 \\ 0 & 1 & 0 \\ -2 & -4 & 9 - \lambda \\ \end{vmatrix} \\ = (1 - \lambda) \begin{vmatrix} 2 - \lambda & -4 \\ -2 & 9 - \lambda \\ \end{vmatrix} \\ = (\lambda - 1)^2(\lambda - 10) \\ 所以，A特征值为1, 1, 10,均为正数,\\ 故,A是正定的.$

$例10.设U是可逆矩阵,且A = U^TU, \\ 证明: f = x^TAx为正定二次型.$
$证:因为U是可逆矩阵, 所以,当x \neq 0时, 有 Ux \neq 0. \\ 事实上, 若Ux = 0, 而U为可逆矩阵, \\ 故, x = U^{-1}Ux = 0,此与x \neq 0矛盾. \\ 而 f = x^T A x = x^TU^TUx = (Ux)^T(Ux) = \Vert Ux \Vert ^2 \\ 又因为 x \neq 0时, Ux \neq 0, 所以有 \Vert Ux \Vert > 0, \\ 故 f > 0 \\ 即 f = x^TAx 为正定二次型.$