跳到主要内容

埼玉大学 理工学研究科 数理電子情報系専攻 情報システム工学コース 2018年8月実施 线性代数

Author

思齐塾, 祭音Myyura

Description

  1. 次の対称行列 AA について考える. [Consider the following symmetric matrix AA .]
A=(301031114)A = \begin{pmatrix} 3 & 0 & -1 \\ 0 & 3 & -1 \\ -1 & -1 & 4 \end{pmatrix}

以下の問に答えよ. [Solve the following problems.]

(a) 対称行列 AA の逆行列を求めよ. [Find the inverse matrix of the symmetric matrix AA .]

(b) 対称行列 AA の固有値と固有ベクトルを求めよ. [Find the eigenvalues and eigenvectors of the symmetric matrix AA .]

(c) 対称行列 AA を対角化するための直交行列 UU を求めよ. また AAUU で対角化したときの対角行列 DD を求めよ. [Find an orthogonal matrix UU for diagonalizing the symmetric matrix AA . Find a diagonal matrix DD by diagonalizing AA with UU .]

(d) 正の整数 nn に対して、 AnA^n を求めよ. [Find AnA^n for a positive integer nn .]

(e) 変数 x,y,zx, y, zx2+y2+z2=1x^2 + y^2 + z^2 = 1 を満たすとき、2次形式 f(x,y,z)=3x2+3y2+4z22yz2zxf(x,y,z) = 3x^2 + 3y^2 + 4z^2 - 2yz - 2zx の最大値と最小値を求めよ. [Find the maximum and minimum of the quadratic form f(x,y,z)=3x2+3y2+4z22yz2zxf(x,y,z) = 3x^2 + 3y^2 + 4z^2 - 2yz - 2zx , if the variables x,yx, y , and zz satisfy x2+y2+z2=1x^2 + y^2 + z^2 = 1 .]

题目描述

  1. 考虑对称矩阵
A=(301031114).A=\begin{pmatrix} 3&0&-1\\ 0&3&-1\\ -1&-1&4 \end{pmatrix}.

回答下列问题。

(a) 求对称矩阵 AA 的逆矩阵。

(b) 求对称矩阵 AA 的特征值与特征向量。

(c) 求一个将对称矩阵 AA 对角化的正交矩阵 UU,并求用 UU 对角化 AA 所得的对角矩阵 DD

(d) 对正整数 nn,求 AnA^n

(e) 当变量 x,y,zx,y,z 满足

x2+y2+z2=1x^2+y^2+z^2=1

时,求二次型

f(x,y,z)=3x2+3y2+4z22yz2zxf(x,y,z)=3x^2+3y^2+4z^2-2yz-2zx

的最大值与最小值。

Kai

(a)

A=(301031114)A = \begin{pmatrix} 3 & 0 & -1 \\ 0 & 3 & -1 \\ -1 & -1 & 4 \end{pmatrix}

det(A)=3(121)0+(1)(0+3)=333=30\det(A) = 3(12-1) - 0 + (-1)(0+3) = 33 - 3 = 30

C=(11131113339)C = \begin{pmatrix} 11 & 1 & 3 \\ 1 & 11 & 3 \\ 3 & 3 & 9 \end{pmatrix}
CT=(11131113339)C^T = \begin{pmatrix} 11 & 1 & 3 \\ 1 & 11 & 3 \\ 3 & 3 & 9 \end{pmatrix}
A1=130(11131113339)A^{-1} = \frac{1}{30} \begin{pmatrix} 11 & 1 & 3 \\ 1 & 11 & 3 \\ 3 & 3 & 9 \end{pmatrix}

(b)

det(AλI)=(3λ)((3λ)(4λ)1)+(1)(0+(3λ))=(3λ)(127λ+λ21)3+λ=(3λ)(λ27λ+11)3+λ=3λ221λ+33λ3+7λ211λ3+λ=λ3+10λ231λ+30=(λ2)(λ3)(λ5)\det(A - \lambda I) = (3-\lambda)((3-\lambda)(4-\lambda) - 1) + (-1)(0 + (3-\lambda)) = (3-\lambda)(12 - 7\lambda + \lambda^2 - 1) - 3 + \lambda = (3-\lambda)(\lambda^2 - 7\lambda + 11) - 3 + \lambda = 3\lambda^2 - 21\lambda + 33 - \lambda^3 + 7\lambda^2 - 11\lambda - 3 + \lambda = -\lambda^3 + 10\lambda^2 - 31\lambda + 30 = -(\lambda - 2)(\lambda - 3)(\lambda - 5)

So the eigenvalues are λ=2,3,5\lambda = 2, 3, 5 .

For λ=2\lambda = 2 : (A2I)v=0(101011112)(xyz)=(000)(A - 2I)v = 0 \Rightarrow \begin{pmatrix} 1 & 0 & -1 \\ 0 & 1 & -1 \\ -1 & -1 & 2 \end{pmatrix} \begin{pmatrix} x \\ y \\ z \end{pmatrix} = \begin{pmatrix} 0 \\ 0 \\ 0 \end{pmatrix} x=z,y=zx = z, y = z . Let z=1z = 1 , then v1=(111)v_1 = \begin{pmatrix} 1 \\ 1 \\ 1 \end{pmatrix} .

For λ=3\lambda = 3 : (A3I)v=0(001001111)(xyz)=(000)(A - 3I)v = 0 \Rightarrow \begin{pmatrix} 0 & 0 & -1 \\ 0 & 0 & -1 \\ -1 & -1 & 1 \end{pmatrix} \begin{pmatrix} x \\ y \\ z \end{pmatrix} = \begin{pmatrix} 0 \\ 0 \\ 0 \end{pmatrix} z=0,x=yz = 0, x = -y . Let y=1y = 1 , then x=1x = -1 , so v2=(110)v_2 = \begin{pmatrix} -1 \\ 1 \\ 0 \end{pmatrix} .

For λ=5\lambda = 5 : (A5I)v=0(201021111)(xyz)=(000)(A - 5I)v = 0 \Rightarrow \begin{pmatrix} -2 & 0 & -1 \\ 0 & -2 & -1 \\ -1 & -1 & -1 \end{pmatrix} \begin{pmatrix} x \\ y \\ z \end{pmatrix} = \begin{pmatrix} 0 \\ 0 \\ 0 \end{pmatrix} 2x=z,2y=z,x+y+z=0x+y2x=0y=x-2x = z, -2y = z, x + y + z = 0 \Rightarrow x + y - 2x = 0 \Rightarrow y = x . 2x=z,x=y-2x = z, x = y . Let x=1x = 1 , then y=1,z=2y = 1, z = -2 , so v3=(112)v_3 = \begin{pmatrix} 1 \\ 1 \\ -2 \end{pmatrix} .

(c) Since A is symmetric, the eigenvectors corresponding to distinct eigenvalues are orthogonal. Thus, U=(13121613121613026)U = \begin{pmatrix} \frac{1}{\sqrt{3}} & -\frac{1}{\sqrt{2}} & \frac{1}{\sqrt{6}} \\ \frac{1}{\sqrt{3}} & \frac{1}{\sqrt{2}} & \frac{1}{\sqrt{6}} \\ \frac{1}{\sqrt{3}} & 0 & -\frac{2}{\sqrt{6}} \end{pmatrix} and D=(200030005)D = \begin{pmatrix} 2 & 0 & 0 \\ 0 & 3 & 0 \\ 0 & 0 & 5 \end{pmatrix} .

(d) Since A=UDUTA = U D U^T , An=UDnUTA^n = U D^n U^T , where Dn=(2n0003n0005n)D^n = \begin{pmatrix} 2^n & 0 & 0 \\ 0 & 3^n & 0 \\ 0 & 0 & 5^n \end{pmatrix} .

An=(13121613121613026)(2n0003n0005n)(13131312120161626)A^n = \begin{pmatrix} \frac{1}{\sqrt{3}} & -\frac{1}{\sqrt{2}} & \frac{1}{\sqrt{6}} \\ \frac{1}{\sqrt{3}} & \frac{1}{\sqrt{2}} & \frac{1}{\sqrt{6}} \\ \frac{1}{\sqrt{3}} & 0 & -\frac{2}{\sqrt{6}} \end{pmatrix} \begin{pmatrix} 2^n & 0 & 0 \\ 0 & 3^n & 0 \\ 0 & 0 & 5^n \end{pmatrix} \begin{pmatrix} \frac{1}{\sqrt{3}} & \frac{1}{\sqrt{3}} & \frac{1}{\sqrt{3}} \\ -\frac{1}{\sqrt{2}} & \frac{1}{\sqrt{2}} & 0 \\ \frac{1}{\sqrt{6}} & \frac{1}{\sqrt{6}} & -\frac{2}{\sqrt{6}} \end{pmatrix}

(e) Let x=13x12y+16zx = \frac{1}{\sqrt{3}} x' - \frac{1}{\sqrt{2}} y' + \frac{1}{\sqrt{6}} z' , y=13x+12y+16zy = \frac{1}{\sqrt{3}} x' + \frac{1}{\sqrt{2}} y' + \frac{1}{\sqrt{6}} z' , z=13x26zz = \frac{1}{\sqrt{3}} x' - \frac{2}{\sqrt{6}} z' . Then 3x2+3y2+4z22yz2zx=2x2+3y2+5z23x^2 + 3y^2 + 4z^2 - 2yz - 2zx = 2x'^2 + 3y'^2 + 5z'^2 . Since x2+y2+z2=1x^2 + y^2 + z^2 = 1 , x2+y2+z2=1x'^2 + y'^2 + z'^2 = 1 . Then 2x2+3y2+5z2=2(x2+y2+z2)+y2+3z2=2+y2+3z22x'^2 + 3y'^2 + 5z'^2 = 2(x'^2 + y'^2 + z'^2) + y'^2 + 3z'^2 = 2 + y'^2 + 3z'^2 . The minimum is 2 (when y=z=0y' = z' = 0 ) and the maximum is 5 (when x=y=0x' = y' = 0 ).