Jephian Lin
    • Create new note
    • Create a note from template
      • Sharing URL Link copied
      • /edit
      • View mode
        • Edit mode
        • View mode
        • Book mode
        • Slide mode
        Edit mode View mode Book mode Slide mode
      • Customize slides
      • Note Permission
      • Read
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Write
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Engagement control Commenting, Suggest edit, Emoji Reply
    • Invite by email
      Invitee

      This note has no invitees

    • Publish Note

      Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

      Your note will be visible on your profile and discoverable by anyone.
      Your note is now live.
      This note is visible on your profile and discoverable online.
      Everyone on the web can find and read all notes of this public team.
      See published notes
      Unpublish note
      Please check the box to agree to the Community Guidelines.
      View profile
    • Commenting
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
      • Everyone
    • Suggest edit
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
    • Emoji Reply
    • Enable
    • Versions and GitHub Sync
    • Note settings
    • Note Insights
    • Engagement control
    • Transfer ownership
    • Delete this note
    • Save as template
    • Insert from template
    • Import from
      • Dropbox
      • Google Drive
      • Gist
      • Clipboard
    • Export to
      • Dropbox
      • Google Drive
      • Gist
    • Download
      • Markdown
      • HTML
      • Raw HTML
Menu Note settings Versions and GitHub Sync Note Insights Sharing URL Create Help
Create Create new note Create a note from template
Menu
Options
Engagement control Transfer ownership Delete this note
Import from
Dropbox Google Drive Gist Clipboard
Export to
Dropbox Google Drive Gist
Download
Markdown HTML Raw HTML
Back
Sharing URL Link copied
/edit
View mode
  • Edit mode
  • View mode
  • Book mode
  • Slide mode
Edit mode View mode Book mode Slide mode
Customize slides
Note Permission
Read
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Write
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Engagement control Commenting, Suggest edit, Emoji Reply
  • Invite by email
    Invitee

    This note has no invitees

  • Publish Note

    Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

    Your note will be visible on your profile and discoverable by anyone.
    Your note is now live.
    This note is visible on your profile and discoverable online.
    Everyone on the web can find and read all notes of this public team.
    See published notes
    Unpublish note
    Please check the box to agree to the Community Guidelines.
    View profile
    Engagement control
    Commenting
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    • Everyone
    Suggest edit
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    Emoji Reply
    Enable
    Import from Dropbox Google Drive Gist Clipboard
       owned this note    owned this note      
    Published Linked with GitHub
    Subscribed
    • Any changes
      Be notified of any changes
    • Mention me
      Be notified of mention me
    • Unsubscribe
    Subscribe
    {%hackmd 5xqeIJ7VRCGBfLtfMi0_IQ %} # 慣性 ![Creative Commons License](https://i.creativecommons.org/l/by/4.0/88x31.png) This work by Jephian Lin is licensed under a [Creative Commons Attribution 4.0 International License](http://creativecommons.org/licenses/by/4.0/). $\newcommand{\trans}{^\top} \newcommand{\adj}{^{\rm adj}} \newcommand{\cof}{^{\rm cof}} \newcommand{\inp}[2]{\left\langle#1,#2\right\rangle} \newcommand{\dunion}{\mathbin{\dot\cup}} \newcommand{\bzero}{\mathbf{0}} \newcommand{\bone}{\mathbf{1}} \newcommand{\ba}{\mathbf{a}} \newcommand{\bb}{\mathbf{b}} \newcommand{\bc}{\mathbf{c}} \newcommand{\bd}{\mathbf{d}} \newcommand{\be}{\mathbf{e}} \newcommand{\bh}{\mathbf{h}} \newcommand{\bp}{\mathbf{p}} \newcommand{\bq}{\mathbf{q}} \newcommand{\br}{\mathbf{r}} \newcommand{\bx}{\mathbf{x}} \newcommand{\by}{\mathbf{y}} \newcommand{\bz}{\mathbf{z}} \newcommand{\bu}{\mathbf{u}} \newcommand{\bv}{\mathbf{v}} \newcommand{\bw}{\mathbf{w}} \newcommand{\tr}{\operatorname{tr}} \newcommand{\nul}{\operatorname{null}} \newcommand{\rank}{\operatorname{rank}} %\newcommand{\ker}{\operatorname{ker}} \newcommand{\range}{\operatorname{range}} \newcommand{\Col}{\operatorname{Col}} \newcommand{\Row}{\operatorname{Row}} \newcommand{\spec}{\operatorname{spec}} \newcommand{\vspan}{\operatorname{span}} \newcommand{\Vol}{\operatorname{Vol}} \newcommand{\sgn}{\operatorname{sgn}} \newcommand{\idmap}{\operatorname{id}} \newcommand{\am}{\operatorname{am}} \newcommand{\gm}{\operatorname{gm}} \newcommand{\iner}{\operatorname{iner}}$ ```python from lingeo import random_int_list from sym import sym_from_list, inertia ``` ## Main idea Let $A$ be an $n\times n$ symmetric matrix. According to the spectral theorem, all eigenvalues of $A$ are real, so they can be arranged from small to large on the real line. Let $n_+(A)$, $n_-(A)$, and $n_0(A)$ be the number of positive, negative, and zero eigenvalues of $A$, respectively. Then the **inertia** of $A$ is defined as $$ \iner(A) = (n_+(A), n_-(A), n_0(A)). $$ Two symmetric matrices $A$ and $B$ are **congruent** if there is an invertible matrix $Q$ such that $$ Q\trans AQ = B. $$ Notice that $Q$ _has to_ be invertible, yet it is $Q\trans$ in the relation. ##### Sylvester's law of inertia If two symmetric matrices are congruent, then they have the same inertia. Moreover, every real symmetric matrix $A$ is congruent to a matrix of the form $$ \begin{bmatrix} I_p & ~ & ~ \\ ~ & -I_q & ~ \\ ~ & ~ & O_r \end{bmatrix}, $$ where $p = n_+(A)$, $q = n_-(A)$, and $r = n_0(A)$. Since every invertible matrix can be decomposed into the product of some elementary matrix. Two symmetric matrices $A$ and $B$ are congruent means there are elementary matrices $E_1,\ldots, E_k$ such that $$ E_k\trans\cdots E_1\trans AE_1\cdots E_k = B. $$ That is, applying some symmetric row/column operations simultaneously to $A$ will result in $B$. ## Side stories - quadratic form - local optimum by derivatives ## Experiments ##### Exercise 1 執行以下程式碼。 ``` set_random_seed(0) print_ans = False n = 3 entries = [1,1] + random_int_list(binomial(n+1,2) - 2, 3) A = sym_from_list(n, entries) pretty_print(LatexExpr("A ="), A) if print_ans: B = copy(A) B.add_multiple_of_row(1,0,-1) B.add_multiple_of_column(1,0,-1) print("A after row/column operation:") show(B) print("(n+, n-, n0) =", inertia(B)) ``` ##### Exercise 1(a) 對 $A$ 進行列運算 $\rho_2:-\rho_1$、再進行行運算 $\kappa_2:-\kappa_1$ 的結果為何? :::warning - [x] seed1 --> `seed = 1` ::: **Ans:** 執行上面的程式碼並使用 `seed = 1` 得到 $$ A = \begin{bmatrix} 1 & 1 & 3 \\ 1 & -1 & 1 \\ 3 & 1 & -2 \\ \end{bmatrix}. $$ 對 $A$進行列運算 $\rho_2:-\rho_1$ 及行運算 $\kappa_2:-\kappa_1$ 得到 $$ A_1 = \begin{bmatrix} 1 & 0 & 3 \\ 0 & -2 & -2 \\ 3 & -2 & -2 \\ \end{bmatrix}. $$ ##### Exercise 1(b) 將 $A$ 進行一系列對稱的行列運算,讓它變成對角矩陣且對角線上只有 $1$、$0$、$-1$。 求 $\iner(A)$。 :::warning - [x] $A_2$ 有錯 ::: **Ans:** 對 $A$ 進行以下對稱的行列運算 $$ \begin{bmatrix} 1 & 0 & 0 \\ -1 & 1 & 0 \\ 0 & 0 & 1 \\ \end{bmatrix} \begin{bmatrix} 1 & 1 & 3 \\ 1 & -1 & 1 \\ 3 & 1 & -2 \\ \end{bmatrix} \begin{bmatrix} 1 & -1 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1 \\ \end{bmatrix} = A_1 = \begin{bmatrix} 1 & 0 & 3 \\ 0 & -2 & -2 \\ 3 & -2 & -2 \\ \end{bmatrix}. $$ $$ \begin{bmatrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ -3 & 0 & 1 \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & 3 \\ 0 & -2 & -2 \\ 3 & -2 & -2 \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & -3 \\ 0 & 1 & 0 \\ 0 & 0 & 1 \\ \end{bmatrix} = A_2 = \begin{bmatrix} 1 & 0 & 0 \\ 0 & -2 & -2 \\ 0 & -2 & -11 \\ \end{bmatrix}. $$ $$ \begin{bmatrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & -1 & 1 \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & 0 \\ 0 & -2 & -2 \\ 0 & -2 & -11 \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & 0 \\ 0 & 1 & -1 \\ 0 & 0 & 1 \\ \end{bmatrix} = A_3 = \begin{bmatrix} 1 & 0 & 0 \\ 0 & -2 & 0 \\ 0 & 0 & -9 \\ \end{bmatrix}. $$ $$ \begin{bmatrix} 1 & 0 & 0 \\ 0 & \frac{2}{\sqrt{2}} & 0 \\ 0 & 0 & 1 \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & 0 \\ 0 & -2 & 0 \\ 0 & 0 & -9 \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & 0 \\ 0 & \frac{2}{\sqrt{2}} & 0 \\ 0 & 0 & 1 \\ \end{bmatrix} = A_4 = \begin{bmatrix} 1 & 0 & 0 \\ 0 & -1 & 0 \\ 0 & 0 & -9 \\ \end{bmatrix}. $$ $$ \begin{bmatrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & \frac{1}{3} \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & 0 \\ 0 & -1 & 0 \\ 0 & 0 & -9 \\ \end{bmatrix} \begin{bmatrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & \frac{1}{3} \\ \end{bmatrix} = A_5 = \begin{bmatrix} 1 & 0 & 0 \\ 0 & -1 & 0 \\ 0 & 0 & -1 \\ \end{bmatrix}. $$ 可求得 $\iner(A) = (1,2,0)$ . ## Exercises ##### Exercise 2 求以下矩陣 $A$ 的 $\iner(A)$。 ##### Exercise 2(a) $$ A = \begin{bmatrix} 1 & 1 \\ 1 & 1 \end{bmatrix}. $$ $Ans:$ $$ \begin{aligned} p_A(x) &= \det(A-xI)\\ &= x^2 - 2x\\ &= x (x - 2),\\ \end{aligned} $$ 當 $p_A(x)=0$,得到 $\spec(A) = \{0,2\}$, 可求得 $\iner(A) = (1,0,1)$。 ##### Exercise 2(b) $$ A = \begin{bmatrix} 0 & 1 \\ 1 & 0 \end{bmatrix}. $$ $Ans:$ $$ \begin{aligned} p_A(x) &= \det(A-xI)\\ &= x^2 - 1\\ &= (x + 1)(x - 1),\\ \end{aligned} $$ 當 $p_A(x)=0$,得到 $\spec(A) = \{-1,1\}$, 可求得 $\iner(A) = (1,1,0)$。 ##### Exercise 2(c) $$ A = \begin{bmatrix} 1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \\ \end{bmatrix}. $$ $Ans:$ $$ \begin{aligned} p_A(x) &= s_0(-x)^3 + s_1(-x)^2 + s_2(-x) + s_3\\ &= (-1)x^3 + 3x^2 \\ &= -x^2(x - 3), \end{aligned} $$ 當 $p_A(x)=0$,得到 $\spec(A) = \{0,0,3\}$, 可求得 $\iner(A) = (1,0,2)$。 ##### Exercise 2(d) $$ A = \begin{bmatrix} 0 & 1 & 1 \\ 1 & 0 & 1 \\ 1 & 1 & 0 \\ \end{bmatrix}. $$ $Ans:$ $$ \begin{aligned} p_A(x) &= s_0(-x)^3 + s_1(-x)^2 + s_2(-x) + s_3\\ &= (-1)x^3 + 3x + 2 \\ &=-(x + 1)(x^2 - x - 2 ) \\ &=-(x + 1)^2(x - 2), \end{aligned} $$ 當 $p_A(x)=0$,得到 $\spec(A) = \{-1,-1,2\}$, 可求得 $\iner(A) = (1,2,0)$。 ##### Exercise 3 一個矩陣 $A$ 的 **二次型(quadratic form)** 指的是長得像 $\bx\trans A\bx$ 的式子。 證明以下關於二次型的性質。 ##### Exercise 3(a) 令 $$ A = \begin{bmatrix} 2 & -1 \\ -1 & 2 \end{bmatrix}, \quad \bx = \begin{bmatrix} x \\ y \end{bmatrix}. $$ 證明 $\bx\trans A\bx \geq 0$。 提示:展開後並將其寫成 $1(ax + by)^2 + 3(cx + dy)^2$。 **Ans:** 我們可將 $\bx\trans A\bx$ 寫成矩陣形式 $$\bx\trans A\bx = \begin{bmatrix} x & y \end{bmatrix} \begin{bmatrix} 2 & -1 \\ -1 & 2 \end{bmatrix} \begin{bmatrix} x \\ y \end{bmatrix} = \begin{bmatrix} 2x^2 - 2xy + 2y^2 \end{bmatrix}. $$ 而 $2x^2-2xy+2y^2$ 又可寫成 $1( \frac{1}{\sqrt{2}} x + \frac{1}{\sqrt{2}}y )^2 + 3( \frac{1}{\sqrt{2}}x + (\frac{-1}{\sqrt{2}})y )^2$。 因為平方為正,所以 **+1** $( \frac{1}{\sqrt{2}} x + \frac{1}{\sqrt{2}}y )^2$ **+3** $( \frac{1}{\sqrt{2}}x + (\frac{-1}{\sqrt{2}})y )^2$ 必為正,因此 $\bx\trans A\bx \geq 0$。 :::success Nice. 不過我後來發覺我的提示有點誤導人,因為 $2x^2 -2xy + 2y^2$ 可以寫成 $(x-y)^2 + x^2 + y^2$ 也是大於等於 $0$。 ::: ##### Exercise 3(b) 令 $A$ 為一 $2\times 2$ 實對稱矩陣,且其特徵值為 $\lambda_1,\lambda_2$。 令 $$ \bx = \begin{bmatrix} x \\ y \end{bmatrix}. $$ 證明 $\bx\trans A\bx$ 可寫成 $\lambda_1(ax + by)^2 + \lambda_2(cx + dy)^2$ 的形式。 :::warning - [x] 寫得不錯,但這題用對稱矩陣可垂直對角化這個性質比較快。而實際上這裡的答案要求對角線上都是 $s$,但對稱矩陣不見得有這個性質。 ::: **Ans:** 因為 $A$ 是實對稱矩陣,所以可以用正交矩陣對角化,假設 $Q$ 是正交矩陣且 $$ Q = \begin{bmatrix} a & c\\ b & d \end{bmatrix}, $$ 那麼 $$ Q^{-1} = Q \trans = \begin{bmatrix} a & b\\ c & d \end{bmatrix}. $$ $Q$ 可以將 $A$ 對角化 $$ Q^{-1} AQ = \begin{bmatrix} a & b\\ c & d \end{bmatrix}A \begin{bmatrix} a &c\\ b & d \end{bmatrix}= \begin{bmatrix} \lambda_1 & 0\\ 0 & \lambda_2 \end{bmatrix}. $$ 因為 $\Col(Q) = \mathbb{R}^2$,$\bx$ 可以表示為 $$ \bx = Q\begin{bmatrix} c_1\\ c_2 \end{bmatrix},\\ c_1,c_2 \in \mathbb{R}. $$ 左右乘上 $Q^{-1}$, $$ \begin{aligned} Q^{-1}\bx &= Q \trans \bx\\ &= \begin{bmatrix} a & b\\ c & d \end{bmatrix} \begin{bmatrix} x\\ y \end{bmatrix}\\ &= \begin{bmatrix} ax+by\\ cx+dy \end{bmatrix}\\ &= \begin{bmatrix} c_1\\ c_2 \end{bmatrix}. \end{aligned} $$ 對 $\bx$ 做轉置, $$ \bx \trans = \begin{bmatrix} c_1 & c_2 \end{bmatrix} Q \trans. $$ 最後考慮 $\bx \trans A \bx$, $$ \begin{aligned} \bx \trans A \bx &= \begin{bmatrix} c_1 & c_2 \end{bmatrix} Q \trans A Q\begin{bmatrix} c_1\\ c_2 \end{bmatrix}\\ &= \begin{bmatrix} c_1 & c_2 \end{bmatrix} \begin{bmatrix} \lambda_1 & 0\\ 0 & \lambda_2 \end{bmatrix} \begin{bmatrix} c_1\\ c_2 \end{bmatrix}\\ &= \lambda_1{c_1}^2 + \lambda_2{c_2}^2\\ &= \lambda_1(ax + by)^2 + \lambda_2(cx + dy)^2. \end{aligned} $$ ##### Exercise 3(c) 令 $A$ 為一 $2\times 2$ 實對稱矩陣,且其特徵值為 $\lambda_1,\lambda_2$。 證明: 1. 若 $\lambda_1, \lambda_2 \geq 0$ 時,$\bx\trans A\bx \geq 0$。 2. 若 $\lambda_1, \lambda_2 \leq 0$ 時,$\bx\trans A\bx \leq 0$。 3. 若 $\lambda_1, \lambda_2 > 0$ 且 $\bx\neq\bzero$ 時,$\bx\trans A\bx > 0$。 4. 若 $\lambda_1, \lambda_2 < 0$ 且 $\bx\neq\bzero$ 時,$\bx\trans A\bx < 0$。 :::warning - [x] 觀念沒錯,不過內容要隨上一題修改 ::: **Ans:** 利用 3(b),$\bx\trans A\bx = \lambda_1(ax + by)^2 + \lambda_2(cx + dy)^2$,3(c) 是顯而易見的. ##### Exercise 4 令 $A$ 為一 $2\times 2$ 實對稱矩陣。 證明: 1. 若 $\det(A) > 0$ 且 $\tr(A) > 0$,則 $\iner(A) = (2,0,0)$。 2. 若 $\det(A) > 0$ 且 $\tr(A) < 0$,則 $\iner(A) = (0,2,0)$。 3. 若 $\det(A) < 0$,則 $\iner(A) = (1,1,0)$。 **Ans:** $A$ 的特徵多項式, $$ p_A(x) = x^2 - \tr(A)x + \det(A). $$ $A$ 的特徵值是特徵多項式的根,所以 $$ \lambda_1 + \lambda_2 = \tr(A)\\ \lambda_1 \lambda_2 = \det(A). $$ 1. 若 $\det(A) > 0$,則 $\lambda_1$ 與 $\lambda_2$ 同號。若 $\tr(A) > 0$,則 $\lambda_1,\lambda_2$ 皆正,所以 $\iner(A) = (2,0,0)$. 2. 若 $\det(A) > 0$,則 $\lambda_1$ 與 $\lambda_2$ 同號。若 $\tr(A) < 0$,則 $\lambda_1,\lambda_2$ 皆負,所以 $\iner(A) = (0,2,0)$. 3. 若 $\det(A) < 0$,則 $\lambda_1$ 與 $\lambda_2$ 異號。所以 $\iner(A) = (1,1,0)$. :::success Great! ::: ##### Exercise 5 令 $f: \mathbb{R}^2 \rightarrow \mathbb{R}$ 為一二次可微函數且其微分連續。 令 $(x_0,y_0)\in\mathbb{R}^2$ 為一點、 而 $f_{xx}, f_{xy} = f_{yx}, f_{yy}$ 分別為 $f$ 對 $x$ 或 $y$ 的二次微分。 令 $$ A = \begin{bmatrix} f_{xx} & f_{yx} \\ f_{xy} & f_{yy} \end{bmatrix}, \quad \bx = \begin{bmatrix} x \\ y \end{bmatrix}. $$ 已知 $f$ 的函數值可以用 $$ f(x_0 + x, y_0 + y) \sim f(x_0, y_0) + \bx\trans A \bx. $$ 逼近。 說明為什麼: 1. 若 $\det(A) > 0$ 且 $\tr(A) > 0$,則 $f$ 在 $(x_0,y_0)$ 有局部最小值。 2. 若 $\det(A) > 0$ 且 $\tr(A) < 0$,則 $f$ 在 $(x_0,y_0)$ 有局部最大值。 3. 若 $\det(A) < 0$,則 $f$ 在 $(x_0,y_0)$ 為𩣑點。 $Ans1:$ 令 $$ x=th ,\ y=tk,\ f(x_0+th,y_0+tk)=g(t),\ f(x_0,y_0)=g(0). $$ 根據全微分公式,我們有 $$ g'(t)= \dfrac{dg}{dt} = \dfrac{\partial f}{\partial x}\cdot\dfrac{dx}{dt}+\dfrac{\partial f}{\partial y}\cdot\dfrac{dy}{dt} = \dfrac{\partial f}{\partial x}\cdot\dfrac{d(x_0+th)}{dt}+\dfrac{\partial f}{\partial y}\cdot\dfrac{d(y_0+tk)}{dt} =\dfrac{\partial f}{\partial x}\cdot h+\dfrac{\partial f}{\partial y}\cdot k, $$ 並且 $$ g''(t)=\dfrac{d}{dt}\Big( \dfrac{\partial f}{\partial x}\cdot h+\dfrac{\partial f}{\partial y}\cdot k \Big) =f_{xx}h^2+2f_{xy}hk+f_{yy}k^2, $$ 以及 $$ g(t)\sim g(0)+g'(0)t+\frac{g''(0)\cdot t^2}{2}. $$ 所以 $$ \Delta f=g(t)-g(0)=\frac{t^2}{2}\cdot (f_{xx}h^2+2f_{xy}hk+f_{yy}k^2). $$ 其中 $$ f_{xx}h^2+2f_{xy}hk+f_{yy}k^2=f_{xx}(h+\frac{k\cdot f_{xy}}{f_{xx}})^2+\frac{f_{xx}\cdot f_{yy}-(f_{xy})^2}{f_{xx}}\cdot k^2. $$ 因為 $\dfrac{t^2}{2}$ 為正, 所以考慮 $f_{xx}h^2+2f_{xy}hk+f_{yy}k^2$ 的正負即可。 若 $\det(A) \gt 0$ 且 $\tr(A) \gt 0$,對於非零向量 $\bf_V=(h,k)$ 可得$$f_{xx}(h+\frac{k\cdot f_{xy}}{f_{xx}})^2+\frac{f_{xx}\cdot f_{yy}-(f_{xy})^2}{f_{xx}}\cdot k^2>0\Rightarrow \Delta\gt 0 \Rightarrow f(x_0+th,y_0+tk)\gt f(x_0,y_0).$$ 所以 $(x_0,y_0)$ 為區域極小值。 若 $\det(A) \gt 0$ 且 $\tr(A) \lt 0$,對於非零向量 $\bf_V=(h,k)$ 可得 $$ f_{xx}(h+\frac{k\cdot f_{xy}}{f_{xx}})^2+\frac{f_{xx}\cdot f_{yy}-(f_{xy})^2}{f_{xx}}\cdot k^2\lt 0\Rightarrow \Delta<0 \Rightarrow f(x_0+th,y_0+tk) \lt f(x_0,y_0). $$ 所以 $(x_0,y_0)$ 為區域極大值。 \ 若 $\det(A) \lt 0$ 且 $f_{xx} \gt 0$, 假設非零向量 $\bf_V=(h,k)=(1,0)$, 則 $$ f_{xx}(h+\frac{k\cdot f_{xy}}{f_{xx}})^2+\frac{f_{xx}\cdot f_yy-(f_{xy})^2}{f_{xx}}\cdot k^2\gt 0, $$ 假設非零向量 $\bf_V=(h,k)=(f_{xy},-f_{xx})$, 則 $$ f_{xx}(h+\frac{k\cdot f_{xy}}{f_{xx}})^2+\frac{f_{xx}\cdot f_{yy}-(f_{xy})^2}{f_{xx}}\cdot k^2\lt 0. $$ 故 $(x_0,y_0)$ 為鞍點。 如果 $\det(A) \lt 0$ 且 $f_{xx} \lt 0$,假設有兩非零向量 $\mathbf{v}=(h,k)=(-f_{xy},f_{xx})$,$(1,0)$。 則我們利用相同推理也可得 $(x_0,y_0)$ 為鞍點。 如果 $f_{xx}=0$,則可以交換 $x,y$ 的位子,得到當 $f_{yy}\neq0$ 時,$(x_0,y_0)$ 為鞍點。 如果 $f_{xx},f_{yy}=0$,由於 $\det\neq0$ 所以 $f_{xy}\neq0$,因此 $f_{xx}h^2+2f_{xy}hk+f_{yy}k^2=2f_{xy}hk{t}^2$。 假設 $f_{xy}\gt 0$, 若依非零向量 $\bf_V=(h,k)=(1,1)$,則 $2f_{xy}hk{t}^2\gt 0$; 若依非零向量 $\bf_V=(h,k)=(1,-1)$,則 $2f_{xy}hk{t}^2\lt 0$。 故 $(x_0,y_0)$ 為鞍點。 假設 $f_{xy}\lt 0$, 若依非零向量 $\bf_V=(h,k)=(1,1)$,則 $f_{xx}h^2+2f_{xy}hk+f_{yy}k^2=2f_{xy}hk{t}^2\lt 0$; 若依非零向量 $\bf_V=(h,k)=(1,-1)$,則 $f_{xx}h^2+2f_{xy}hk+f_{yy}k^2=2f_{xy}hk{t}^2\gt 0$。 故 $(x_0,y_0)$ 為鞍點。 根據上述推理,因此題目敘述得證。 **Ans2:** 利用Exercise 4的結果, 1. 若 $\det(A) > 0$ 且 $\tr(A) > 0$,則 $$ f(x_0 + x, y_0 + y) - f(x_0, y_0) \sim \bx\trans A \bx \geq 0. $$ $f$ 在 $(x_0,y_0)$ 附近都不比在 $(x_0,y_0)$ 小,所以 $f$ 在$(x_0,y_0)$ 有局部最小值。 2. 若 $\det(A) > 0$ 且 $\tr(A) < 0$,則 $$ f(x_0 + x, y_0 + y) - f(x_0, y_0) \sim \bx\trans A \bx \leq 0. $$ $f$ 在 $(x_0,y_0)$ 附近都不比在 $(x_0,y_0)$ 大,所以 $f$ 在$(x_0,y_0)$ 有局部最大值。 3. 若 $\det(A) < 0$,則可以取適當的 $\bx$ 使得 $\bx\trans A \bx \geq 0$ 或 $\bx\trans A \bx \leq 0$, 所以 $f$ 在 $(x_0,y_0)$ 是鞍點。 ##### Exercise 6 利用以下步驟證明西爾維斯特慣性定理。 ##### Sylvester's law of inertia If two symmetric matrices are congruent, then they have the same inertia. ##### Exercise 6(a) 定義 $$ E(t) = \begin{bmatrix} 1 & t \\ 0 & 1 \end{bmatrix}. $$ 說明 $E(t)\trans AE(t)$ 在任何 $t\in\mathbb{R}$ 時都有相同的零維數。 :::warning 這題主要是說明如果 $E$ 可逆,則 $\nul(E\trans AE) = \nul(A)$ ::: **Ans:** 回憶一下, $$ \rank(ABC) \leq \min(\rank(A), \rank(B), \rank(C)),\\ \rank(A) + \nul(A) = n. $$ 因為 $E(t)$ 對於所有 $t \in \mathbb{R}$ 都是滿秩的,所以 $\rank(E(t)) = 2$,且 $E(t)$ 可逆。 令 $$ \begin{aligned} E(t) \trans AE(t) &= B,\\ (E(t) \trans)^{-1} B (E(t))^{-1} &= A. \end{aligned} $$ 因為 $E(t)$ 是滿秩的,所以 $\rank(A), \rank(B) \leq \rank(E(t)) = \rank(E(t) \trans)$ 那麼 $$ \begin{aligned} &\rank(B) \leq \min(\rank(E(t) \trans), \rank(A), \rank(E(t))) = \rank(A)\\ &\rank(A) \leq \min(\rank(E(t) \trans)^{-1}, \rank(B), \rank((E(t))^{-1})) = \rank(B)\\ \end{aligned} $$ 所以 $$ \begin{aligned} \rank(A) &= \rank(B)\\ &= \rank(E(t) \trans AE(t)),\\ n - \rank(A) &= \nul(A)\\ &= \nul(E(t) \trans AE(t))\\ &= n - \rank(E(t) \trans AE(t)). \end{aligned} $$ ##### Exercise 6(b) 定義 $$ E(t) = \begin{bmatrix} t & 0 \\ 0 & 1 \end{bmatrix}. $$ 說明 $E(t)\trans AE(t)$ 在任何 $t > 0$ 時都有相同的零維數。 **Ans:** 因為對於任何 $t > 0$,$E(t)$ 都是滿秩的,根據6(a), $$ \nul(E(t) \trans AE(t)) = \nul(A). $$ ##### Exercise 6(c) 定義 $$ E = \begin{bmatrix} -1 & 0 \\ 0 & 1 \end{bmatrix}. $$ 說明 $\iner(E\trans AE) = \iner(A)$。 **Ans:** 令 $$ A = \begin{bmatrix} a & b\\ c & d \end{bmatrix}. $$ 直接計算 $E \trans AE$, $$ E \trans AE = \begin{bmatrix} a & -b\\ -c & d \end{bmatrix}. $$ 因為 $$ \begin{aligned} \det(A) &= \det(E \trans AE) = ad - bc\\ \tr(A) &= \tr(E \trans AE) = a + d. \end{aligned} $$ 還有 Exercise 4 的結果,所以 $\iner(E \trans AE) = \iner(A)$. ##### Exercise 6(d) 已知矩陣的特徵值會隨矩陣的數值連續變動。 利用這個性質證明西爾維斯特慣性定理。 :::warning - [ ] 這題要說明 $F$ 要怎麼把 $I$ 變成 $Q$ - [ ] 不可逆的部份論證有問題;說明在變化過程中 $0$ 的個數不變就好 ::: **Ans:** 若一個矩陣是可逆的,則該矩陣的 $\ker = \{ \bzero \}$,也就是該矩陣特徵值不為 $0$. 若 $A$ 可逆,則 $Q \trans AQ = B$ 也是可逆的,所以 $A,B$ 的特徵值皆不為 $0$. 定義一個連續可逆函數 $F : [0,1] \rightarrow M_{n \times n}$,且 $F(0) = I_n$,$F(1) = Q$. 當 $t$ 從 $0$ 增大到 $1$,$F(t) \trans A F(t)$ 的特徵值慢慢從 $A$ 的特徵值變成 $B$ 的特徵值,因為 $F(t) \trans A F(t)$ 一直都是可逆的,所以特徵值變化的過程都沒有經過 $0$,所以 $\iner(A) = \iner(B)$. 若 $A$ 不可逆,考慮可逆矩陣 $A + \epsilon I$, $$ \begin{aligned} Q \trans (A + \epsilon I) Q &= Q \trans AQ + \epsilon Q \trans Q\\ &= B + \epsilon Q \trans Q. \end{aligned} $$ 根據上一個結論,所以 $\iner(A + \epsilon I) = \iner(B + \epsilon Q \trans)$. 當 $\epsilon \rightarrow 0$,$\iner(A) = \iner(B)$. :::info 分數 = 7 :::

    Import from clipboard

    Paste your markdown or webpage here...

    Advanced permission required

    Your current role can only read. Ask the system administrator to acquire write and comment permission.

    This team is disabled

    Sorry, this team is disabled. You can't edit this note.

    This note is locked

    Sorry, only owner can edit this note.

    Reach the limit

    Sorry, you've reached the max length this note can be.
    Please reduce the content or divide it to more notes, thank you!

    Import from Gist

    Import from Snippet

    or

    Export to Snippet

    Are you sure?

    Do you really want to delete this note?
    All users will lose their connection.

    Create a note from template

    Create a note from template

    Oops...
    This template has been removed or transferred.
    Upgrade
    All
    • All
    • Team
    No template.

    Create a template

    Upgrade

    Delete template

    Do you really want to delete this template?
    Turn this template into a regular note and keep its content, versions, and comments.

    This page need refresh

    You have an incompatible client version.
    Refresh to update.
    New version available!
    See releases notes here
    Refresh to enjoy new features.
    Your user state has changed.
    Refresh to load new user state.

    Sign in

    Forgot password

    or

    By clicking below, you agree to our terms of service.

    Sign in via Facebook Sign in via Twitter Sign in via GitHub Sign in via Dropbox Sign in with Wallet
    Wallet ( )
    Connect another wallet

    New to HackMD? Sign up

    Help

    • English
    • 中文
    • Français
    • Deutsch
    • 日本語
    • Español
    • Català
    • Ελληνικά
    • Português
    • italiano
    • Türkçe
    • Русский
    • Nederlands
    • hrvatski jezik
    • język polski
    • Українська
    • हिन्दी
    • svenska
    • Esperanto
    • dansk

    Documents

    Help & Tutorial

    How to use Book mode

    Slide Example

    API Docs

    Edit in VSCode

    Install browser extension

    Contacts

    Feedback

    Discord

    Send us email

    Resources

    Releases

    Pricing

    Blog

    Policy

    Terms

    Privacy

    Cheatsheet

    Syntax Example Reference
    # Header Header 基本排版
    - Unordered List
    • Unordered List
    1. Ordered List
    1. Ordered List
    - [ ] Todo List
    • Todo List
    > Blockquote
    Blockquote
    **Bold font** Bold font
    *Italics font* Italics font
    ~~Strikethrough~~ Strikethrough
    19^th^ 19th
    H~2~O H2O
    ++Inserted text++ Inserted text
    ==Marked text== Marked text
    [link text](https:// "title") Link
    ![image alt](https:// "title") Image
    `Code` Code 在筆記中貼入程式碼
    ```javascript
    var i = 0;
    ```
    var i = 0;
    :smile: :smile: Emoji list
    {%youtube youtube_id %} Externals
    $L^aT_eX$ LaTeX
    :::info
    This is a alert area.
    :::

    This is a alert area.

    Versions and GitHub Sync
    Get Full History Access

    • Edit version name
    • Delete

    revision author avatar     named on  

    More Less

    Note content is identical to the latest version.
    Compare
      Choose a version
      No search result
      Version not found
    Sign in to link this note to GitHub
    Learn more
    This note is not linked with GitHub
     

    Feedback

    Submission failed, please try again

    Thanks for your support.

    On a scale of 0-10, how likely is it that you would recommend HackMD to your friends, family or business associates?

    Please give us some advice and help us improve HackMD.

     

    Thanks for your feedback

    Remove version name

    Do you want to remove this version name and description?

    Transfer ownership

    Transfer to
      Warning: is a public team. If you transfer note to this team, everyone on the web can find and read this note.

        Link with GitHub

        Please authorize HackMD on GitHub
        • Please sign in to GitHub and install the HackMD app on your GitHub repo.
        • HackMD links with GitHub through a GitHub App. You can choose which repo to install our App.
        Learn more  Sign in to GitHub

        Push the note to GitHub Push to GitHub Pull a file from GitHub

          Authorize again
         

        Choose which file to push to

        Select repo
        Refresh Authorize more repos
        Select branch
        Select file
        Select branch
        Choose version(s) to push
        • Save a new version and push
        • Choose from existing versions
        Include title and tags
        Available push count

        Pull from GitHub

         
        File from GitHub
        File from HackMD

        GitHub Link Settings

        File linked

        Linked by
        File path
        Last synced branch
        Available push count

        Danger Zone

        Unlink
        You will no longer receive notification when GitHub file changes after unlink.

        Syncing

        Push failed

        Push successfully