new post

Bump hexo-word-counter from 0.2.0 to 0.2.1

_config.yml

@@ -109,9 +109,6 @@ theme: next
 # Deployment
 ## Docs: https://hexo.io/docs/one-command-deployment
 deploy:
-  - type: git
-    repo: [email protected]:/home/josh/documents/josh-blog-deploy.git
-    branch: master
   - type: git
     repo: [email protected]:/home/josh/documents/josh-blog-deploy.git
     branch: master
@@ -119,12 +116,6 @@ deploy:
     repo: https://github.com/joshgao22/joshgao22.github.io.git
     branch: master
 
-# MathJax [Redundant]
-# math:
-#   engine: 'mathjax'
-#   mathjax:
-#     src: https://cdnjs.cloudflare.com/ajax/libs/mathjax/3.2.2/es5/tex-mml-chtml.min.js
-
 # Local Search
 search:
   path: search.xml
@@ -198,12 +189,18 @@ neat_js:
 
 related_posts:
   enabled: true
+  enable_env_name: prod
   filter_threshold: 0.13
   related_count: 5
   weight:
     title: 0.1
-    description: 0
-    keywords: 0
-    tags: 0.01
-    categories: 0.05
-    text: 0
+    description: 0.01
+    keywords: 0.01
+    tags: 0.005
+    categories: 0.005
+    text: 0.2
+
+# config of hexo-renderer-pandoc
+pandoc:
+  args:
+    - --mathjax

package.json

@@ -12,7 +12,7 @@
     "version": "7.3.0"
   },
   "dependencies": {
-    "@next-theme/plugins": "^8.23.0",
+    "@next-theme/plugins": "^8.23.1",
     "hexo": "^7.3.0",
     "hexo-abbrlink": "^2.2.1",
     "hexo-deployer-git": "^4.0.0",
@@ -24,16 +24,16 @@
     "hexo-generator-sitemap": "^3.0.1",
     "hexo-generator-tag": "^2.0.0",
     "hexo-helper-live2d": "^3.1.1",
-    "hexo-markmap": "^1.2.5",
+    "hexo-markmap": "^2.0.8",
     "hexo-neat": "^1.0.9",
     "hexo-related-posts": "^1.6.2",
     "hexo-renderer-ejs": "^2.0.0",
-    "hexo-renderer-pandoc": "^0.4.0",
+    "hexo-renderer-pandoc": "^0.5.0",
     "hexo-renderer-stylus": "^3.0.1",
     "hexo-server": "^3.0.0",
     "hexo-tag-cloud": "2.1.*",
     "hexo-theme-landscape": "^1.1.0",
-    "hexo-theme-next": "^8.23.0",
-    "hexo-word-counter": "^0.2.0"
+    "hexo-theme-next": "^8.23.1",
+    "hexo-word-counter": "^0.2.1"
   }
 }

source/_drafts/2025-06-26-josh-adv-math-1-2.md

@@ -0,0 +1,35 @@
+---
+title: Josh's Summary — 工科数学分析<br>Part 1.2 常微分方程
+mathjax: true
+comments: true
+copyright: true
+toc:
+  enable: true
+  number: true
+  wrap: true
+  expand_all: true
+  max_depth: 4
+tags:
+  - 高等数学
+  - 数学分析
+  - 微积分
+categories:
+  - - Josh 的学习笔记
+    - 高等数学
+  - - 数学
+    - 数学分析
+abbrlink: 5a263f58
+date: 2025-06-26 10:21:02
+---
+
+# 常见等价无穷小
+
+1. $\sin x \sim x$
+2. $\tan x \sim x$
+3. $\arcsin x \sim x$
+4. $\arctan x \sim x$
+5. $1 - \cos x \sim \frac{x^2}{2}$
+6. $\ln(1+x) \sim x$
+7. $e^x - 1 \sim x$
+8. $a^x - 1 \sim x \ln a$
+9. $(1+x)^\alpha - 1 \sim \alpha x$

source/_posts/2023-04-14-calculas-formulas.md

@@ -1,194 +0,0 @@
----
-title: 高等数学常用公式
-mathjax: true
-comments: true
-copyright: true
-toc:
-  enable: true
-  number: false
-  wrap: true
-  expand_all: true
-  max_depth: 4
-abbrlink: 1823c9d1
-date: 2023-04-14 16:25:59
-tags:
-  - 等价无穷小
-  - 导数公式
-  - Maclaurin 公式
-  - 曲率
-  - 积分公式
-  - 三角函数
-categories:
-  - [数学, 微积分]
----
-
-# 1. 常见等价无穷小
-
-||||
-| :--: | :--: | :--: |
-| $\sin x \sim x$ | $\tan x \sim x$ | $\arcsin x \sim x$ |
-| $\arctan x \sim x$ | $\displaystyle\left( 1 - \cos x \right) \sim \frac{x^2}{2}$ | $\ln (1+x) \sim x$ |
-| $\left( e^x -1 \right) \sim x$ | $\left( a^x - 1 \right) \sim x \ln a$ | $\left( \left( 1+x \right) ^\alpha -1 \right) \sim \alpha x$  |
-
-<!-- more -->
-
-# 2. 导数基本公式
-
-| 一般函数 | 三角函数 | 反三角函数 | 双曲函数<br>(圆函数) |
-| :--: | :--: | :--: | :----: |
-| $\left( C \right)' = 0$ | $\left( \sin x \right)' = \cos x$ | $\displaystyle\left( \arcsin x \right)' = \frac{1}{\sqrt{1-x^2}}$ | $\displaystyle\left( \sinh x \right)' = \left(\frac{e^x-e^{-x}}{2}\right)' = \cosh x$ |
-| $\left( x^\alpha \right)'=\alpha x^{\alpha-1} \left( \alpha \in  \mathbf{Z} \right)$ | $\left( \cos x \right)' = -\sin x$ | $\displaystyle\left( \arccos x \right)' = -\frac{1}{\sqrt{1-x^2}}$ | $\displaystyle\left( \cosh x \right)' = \left( \frac{e^x+e^{-x}}{2} \right)'=\sinh x$ |
-| $\displaystyle\left( \log_ax\right)'=\frac{1}{x\ln a}$ | $\displaystyle\left( \tan x \right)' = \frac{1}{\cos^2 x}=\sec^2 x$ | $\displaystyle\left( \arctan x \right)' = \frac{1}{1+x^2}$ | $\displaystyle\left( \mathrm{arsinh} \, x \right)' = \left(\ln \left( x+\sqrt{x^2 + 1} \right) \right)' = \frac{1}{x^2+1}$ |
-| $( \ln$\| $x$ \|$)'=\displaystyle\frac{1}{x}$ | $\displaystyle\left( \cot x \right)' = -\frac{1}{\sin^2 x} = -\csc^2 x$ | $\displaystyle\left( \mathrm{arccot} \, x \right)' = -\frac{1}{1+x^2}$ | $\displaystyle\left( \mathrm{arcosh} \, x \right)' = \left( \ln \left( x + \sqrt{x^2 - 1} \right) \right)' = \frac{1}{\sqrt{x^2-1}}$
-|  | $\left( \sec x \right)' = \sec x \cdot \tan x$ |
-|  | $\left( \csc x \right)' = \csc x \cdot \cot x$ |
-
-- Newton-Leibniz 公式：
-
-  $$
-  \begin{aligned}
-  \left( u \cdot v \right)^{\left( n \right)} &= u^{\left( n \right)}v + nu^{\left( n-1 \right)}v' + \frac{n \left( n-1 \right)}{2!}u^{\left( n-2 \right)}v'' + \cdots + \frac{n\left( n-1 \right)\cdots\left(n-k+1\right)}{k!}u^{\left( n-k \right)}v^{(k)} + \cdots + uv^{\left( n \right)} \\
-  &= \sum_{k=0}^{n}C_n^k u^{\left( u-k \right)}v^k
-  \end{aligned}
-  $$
-
-  可以注意到 Newton-Leibniz 公式具有类似二项式定理展开的形式。
-
-# 3. 几个初等函数的 Maclaurin 公式
-
-## 3.1 Taylor 公式的一般形式
-
-$$
-f(x)=f(x_0) + f'(x_0)(x-x_0) + \frac{f''(x_0)}{2!}(x-x_0)^2 + \cdots + \frac{f^{(n)}(x_0)}{n!}(x-x_0)^n + R_n
-$$
-
-其中
-
-$$
-R_n = \frac{f^{(n+1)}(\xi)}{(n+1)!}(x-x_0)^{n+1} \, (\xi\,在\,x\,和\,x_0\,之间) = o\left( \left( x - x_0 \right)^n \right)(x\to x_0)
-$$
-
-## 3.2 Maclaurin 公式的一般形式
-
-$$
-f(x)=f(0)+f'(0)\,x + \frac{f''(0)}{2!}x^2 + \cdots + \frac{f^{(n)}(0)}{n!}x^n + R_n
-$$
-
-其中
-
-$$
-R_n = \frac{f^{(n+1)}(\xi)}{(n+1)!}x^{n+1} \, (\xi\,在\,0\,和\,x\,之间) = \frac{f^{(n+1)}(\theta x)}{(n+1)!} x^{n+1}(0<\theta<1) = o\left( \left( x - x_0 \right)^n \right)(x\to x_0)
-$$
-
-## 3.3 指数函数的 Maclaurin 展开式
-
-$$
-e^x=1+x+\frac{x^2}{2} + \frac{x^3}{3} + \cdots +\frac{x^n}{n!} + \frac{e^{\theta x}}{(n+1)!}(0<\theta<1)
-$$
-
-## 3.4 正弦函数的 Maclaurin 展开式
-
-$$
-\sin x = x - \frac{x^3}{3!} + \frac{x^5}{5!} + \cdots + (-1)^{m-1}\frac{x^{2m-1}}{(2m-1)!} + R_{2m}
-$$
-
-其中
-
-$$
-R_{2m}=\frac{\sin\left( \theta x + \frac{(2m-1)\pi}{2}\right)}{(2m+1)!}x^{2m+1} = o(x^{2m})(0<\theta <1)
-$$
-
-## 3.5 余弦函数的 Maclaurin 展开式
-
-$$
-\cos x = 1 - \frac{x^2}{2!} + \frac{x^4}{4!} + \cdots + (-1)^m\frac{x^{2m}}{2m!} + R_{2m+1}
-$$
-
-其中
-
-$$
-R_{2m+1}=\frac{\cos(\theta x + (m+1)\pi)}{(2m+2)!} x^{2m+2}=o(x^{2m+2}) (0<\theta <1)
-$$
-
-## 3.6 $\ln (1+x)$ 的 Maclaurin 展开式
-
-$$
-\ln(1+x) = x - \frac{x^2}{2} + \frac{x^3}{3} -\frac{x^4}{4} + \cdots + \frac{(-1)^{n-1}x^n}{n} + R_n(x)
-$$
-
-其中
-
-$$
-R_n(x) = \frac{(-1)^n}{(n+1)(1+\theta x)^{n+1}}x^{n+1} = o(x^n)(0<\theta <1)
-$$
-
-## 3.7 $(1+x)^\alpha$ 的 Maclaurin 展开式
-
-$$
-(1+x)^\alpha = 1 + \alpha x + \frac{\alpha(\alpha-1)}{2!}x^2 + \cdots + \frac{\alpha(\alpha - 1)\cdots (\alpha-n+1)}{n!}x^n + R_n(x)
-$$
-
-其中
-
-$$
-R_n(x) = \frac{\alpha(\alpha -1)\cdots(\alpha-n)}{(n+1)!}(1+\theta x)^{\alpha-n-1}x^{n+1} = o(x)(0<\theta <1)
-$$
-
-## 3.8 $\frac{1}{1-x}$ 的 Maclaurin 展开式
-
-$$
-\frac{1}{1-x} = 1+x+x^2+\cdots+x^n +o(x^n)
-$$
-
-# 4. 曲率及曲率半径
-
-| 曲率 | 曲率半径 |
-| :--: | :--: |
-| $K=\left\| \dfrac{\mathrm{d}\alpha}{\mathrm{d}s}\right\| \xlongequal{\tan \alpha = y' \Rightarrow \alpha = \arctan y'} \dfrac{\left\| \frac{y''}{1+y'^2}\mathrm{d}x \right\|}{\sqrt{1+y'^2}}\mathrm{d}x = \dfrac{\left\| y'' \right\|}{\left( 1 + y' \right)^\frac{3}{2}}$ | $R = \frac{1}{K}$ |
-
-# 5. 积分常用公式
-
-| 序 | 描述 | 公式 |
-| :--: | :----: | :----------------------------------------------:|
-| 1 | 定积分的定义 | $\displaystyle\int^1_0f(x)\mathrm{d}x=I=\lim_{\lambda \to 0}\sum_{i=1}^{n}f(\xi_i)\Delta x_i$，其中 $\lambda = \max \left\{ \Delta x_1,\Delta x_2, \cdots, \Delta x_n \right\}$ |
-| 2 | 变上限积分 | $考察定积分\displaystyle\int^x_af(x)\mathrm{d}x=\int^x_af(t)\mathrm{d}t$，对 $\displaystyle\Phi(x)=\int_a^xf(t)\mathrm{d}t$，有  $\displaystyle\Phi'(x)=\frac{\mathrm{d}}{\mathrm{d}x}\int_a^xf(t)\mathrm{d}t = f(x)$ |
-| 3. | — | $\displaystyle\frac{\mathrm{d}}{\mathrm{d}x}\int^{\beta(x)}_{\alpha(x)}f'(t)\mathrm{d}t = f\left[ \beta\left(x\right)\right]\beta'(x)-f[\alpha(x)]\alpha'(x)$ |
-| 4 | 上下限对称的积分 | 若 $f(x)$ 为偶函数，则 $\displaystyle\int_{-a}^af(x)\mathrm{d}x=2\int_0^af(x)\mathrm{d}x$；若 $f(x)$ 为奇函数，则 $\displaystyle\int_{-a}^af(x)\mathrm{d}x=0$ |
-| 5 | 周期函数的积分 | $\displaystyle f(x+T)=f(x)\Rightarrow\int_0^{nT}f(x)\mathrm{d}x=n\int^T_0f(x)\mathrm{d}x$ |
-| 6 | 分部积分法 | $\displaystyle\int uv'\mathrm{d}x=uv - \int u'v\mathrm{d}x \Leftrightarrow \int u\mathrm{d}v = uv - \int v\mathrm{d}u$ |
-
-# 6. 基本积分表
-
-| 序 | 积分式 |
-| :--: | :-------------------------: |
-| 1 | $\displaystyle\int k\mathrm{d}x=kx+C$，特别地 $\displaystyle\int\mathrm{d}x=x+C$ |
-| 2 | $\displaystyle\int x^\mu\mathrm{d}x=\frac{x^{\mu+1}}{\mu+1}+C(\mu\ne-1)$，特别地 $\displaystyle\int \frac{\mathrm{d}x}{x^2}=-\frac{1}{x}+C,\int\frac{\mathrm{d}x}{\sqrt{x}}=2\sqrt{x}+C$ |
-| 3 | $\displaystyle\int\frac{\mathrm{d}x}{x}=\ln \|x\|+C$ |
-| 4 | $\displaystyle\int\frac{\mathrm{d}x}{1+x^2} = \arctan x +C$ |
-| 5 | $\displaystyle\int\frac{\mathrm{d}x}{\sqrt{1-x^2}}=\arcsin x+C$ |
-| 6 | $\displaystyle\int\sin x\mathrm{d}x = -\cos x+C, \int\cos x\mathrm{d}x = \sin x +C$ |
-| 7 | $\displaystyle\int\frac{\mathrm{d}x}{\cos^2x} = \int\sec^2x\mathrm{d}x = \tan x +C, \int\frac{\mathrm{d}x}{\sin^2x}=\int\csc^2x\mathrm{d}x = -\cot x +C$ |
-| 8 |  $\displaystyle\int a^x\mathrm{d}x = \frac{a^x}{\ln a} + C$，特别地 $\displaystyle\int e^x\mathrm{d}x = e^x +C$
-| 9 | $\displaystyle\int \sinh x\mathrm{d}x = \cosh x + C，\int \cosh x \mathrm{d}x=\sinh x +C$ |
-
-# 7. 三角函数相关公式
-
-| 序 | 公式 |
-| :--: | :-------------------------: |
-| 1 | $\sin^2x+\cos^2x=1, \tan^2x+1=\sec^2x,\cot^2x+1=\csc^2x$ |
-| 2 | $\begin{aligned}&\cosh^2x-\sinh^2x=1,\\ &\sinh(x\pm y)=\sinh x\cdot\cosh y \pm \sinh y\cdot\cosh x, \cosh(x\pm y) = \cosh x\cdot\cosh y,\\ &\sinh2x = 2\sinh x\cdot \cosh x, \cosh 2x = \cosh^2x+\sinh^2x = 2\cosh^2x - 1\end{aligned}$ |
-| 3<br>积化和差公式 | $\sin\alpha\cos\beta=\frac12[\sin(\alpha+\beta)+\sin(\alpha-\beta)],\cos\alpha\sin\beta = \frac12[\sin(\alpha+\beta) - \sin(\alpha-\beta)]$ <br> $\cos\alpha\cos\beta = \frac12[cos(\alpha+\beta) + cos(\alpha - \beta)], \sin\alpha\sin\beta=-\frac12[\cos(\alpha+\beta) - \cos(\alpha-\beta)]$ |
-| 4<br>和差化积公式 | $\displaystyle\sin\alpha + \sin\beta = 2\sin\left(\frac{\alpha+\beta}{2}\right) \cos\left(\frac{\alpha-\beta}{2}\right), \sin\alpha - \sin\beta = 2\sin\left(\frac{\alpha-\beta}{2}\right) \cos\left(\frac{\alpha+\beta}{2}\right)$ <br> $\displaystyle\cos\alpha + \cos\beta = 2\cos\left(\frac{\alpha+\beta}{2}\right) \cos\left(\frac{\alpha-\beta}{2}\right), \cos\alpha - \cos\beta = -2\sin\left(\frac{\alpha+\beta}{2}\right) \sin\left(\frac{\alpha-\beta}{2}\right)$ |
-| 5<br>三角代换 | $\displaystyle\sqrt{a^2-x^2}\xrightarrow{x=a\sin t}a\cos t, \sqrt{a^2+x^2} \xrightarrow{x=a\tan t}a\sec t, \sqrt{x^2-a^2}\xrightarrow{x=a\sec t}a\tan t \left( a>0,t\in\left( -\frac{\pi}{2} , \frac{\pi}{2} \right) \right)$ |
-| 6 | $\displaystyle\int\sqrt{a^2-x^2}\mathrm{d}x = \frac{a^2}{2}\arcsin\frac xa + \frac12x\sqrt{a^2-x^2} + C(a>0)$ |
-| 7 | $\displaystyle\int\sqrt{x^2+a^2}\mathrm{d}x = \frac x2\sqrt{x^2+a^2} + \frac{a^2}2\ln \left\| x+\sqrt{x^2+a^2} \right\|+C(a>0)$ |
-| 8 | $\displaystyle\int\frac{1}{\sqrt{x^2+a^2}}\mathrm{d}x = \ln\left\| x+ \sqrt{x^2+a^2} \right\| +C$, $\displaystyle\int\frac{1}{\sqrt{x^2-a^2}}\mathrm{d}x = \ln\left\| x + \sqrt{x^2+a^2}\right\|(a>0)$，特别地，$\displaystyle\int\frac{1}{\sqrt{x^2+1}}\mathrm{d}x=\left(x+\sqrt{x^2+1}\right) = \mathrm{arsinh}\, x$, $\displaystyle\int\frac{1}{\sqrt{x^2-1}}\mathrm{d}x = \ln\left( x+\sqrt{x^2-1}\right) = \mathrm{arcosh}\, x$ |
-| 9 | $\displaystyle\int\sec x\mathrm{d}x = \int\frac{1}{\cos x}\mathrm{d}x = \int\frac{\sec^2x+\sec x \tan x}{\sec x + \tan x}\mathrm{d}x = \int\frac{1}{\sec x + \tan x}\mathrm{d}(\sec x + \tan x) = \ln \left\| \sec x + \tan x \right\| + C$ |
-| 10 | $\displaystyle\int\csc x\mathrm{d}x = \int\frac{1}{\sin x}\mathrm{d}x = \ln\left\| \tan\frac x2 \right\| + C = \frac 12 \ln \left\| \frac{1-\cos x}{1 + \cos x}\right\| + C$ |
-| 11 | $\displaystyle\int\tan x\mathrm{d}x = \int\frac{\sin x}{\cos x}\mathrm{d}x = -\int\frac{1}{\cos x}\mathrm{d}(\cos x) = -\ln \left\| \cos x\right\| +C$ |
-| 12 | $\displaystyle\int\frac{1}{\tan x}\mathrm{d}x = \int\frac{\cos x}{\sin x}\mathrm{d}x = \int\frac{1}{\sin x}\mathrm{d}(\sin x) = \ln\left\| \sin x \right\| + C$ |
-| 13 | $\displaystyle\int\arctan x\mathrm{d}x = x\arctan x - \ln \left\| 1+x^2 \right\| +C$ |
-| 14<br>Wallis 公式 | $\displaystyle\int^{\frac{\pi}{2}}_{0}\sin^nx\mathrm{d}x = \int^{\frac{\pi}{2}}_{0}\cos^nx\mathrm{d}x = \begin{cases} \displaystyle\frac{n-1}{n}\cdot\frac{n-3}{n-2}\cdot\cdots\cdot\frac12\cdot\frac\pi2,&n为偶数 \\ \displaystyle\frac{n-1}{n}\cdot\frac{n-3}{n-2}\cdot\cdots\cdot\frac23\cdot1,&n为奇数 \end{cases} \\ = \begin{cases} \displaystyle\frac{(n-1)!!}{n!!}\cdot\frac\pi2,&n为偶数 \\ \displaystyle\frac{(n-1)!!}{n!!},&n为奇数 \end{cases}$ |
-| 15 | $\displaystyle\int_0^\pi\sin^nx\mathrm{d}x = 2\int^\frac\pi2_0\sin^nx\mathrm{d}x,\int_0^\pi\cos^nx\mathrm{d}x = \begin{cases} 2\displaystyle\int^\frac\pi2_0\cos^nx\mathrm{d}x,&n为偶数 \\ 0,&n为奇数 \end{cases}$ <br> $\displaystyle\int_0^{2\pi}\sin^nx = \int_0^{2\pi}\cos^nx\mathrm{d}x = \begin{cases} 4\displaystyle\int^\frac\pi2_0\sin^nx\mathrm{d}x,&n为偶数 \\ 0,&n为奇数 \end{cases}$ <br> $\displaystyle\int_0^\frac\pi2f(\sin x)\mathrm{d}x = \int_0^\frac\pi2f(\cos x)\mathrm{d}x$, $\displaystyle\int_0^\pi f(\sin x)\mathrm{d}x \ne \int_0^\pi f(\cos x)\mathrm{d}x$ <br> $\displaystyle\int_0^\pi\pi f(\sin x)\mathrm{d}x = \frac\pi2\int_0^\pi f(\sin x)\mathrm{d}x = \pi\int_0^\frac\pi2 f(\sin x)\mathrm{d}x$ |
-| 16<br>万能公式 | 令 $u=\displaystyle\tan\frac x2,x=2\arctan u$，则 $\displaystyle\sin x = \frac{2u}{1+u^2},\cos x=\frac{1-u^2}{1+u^2}, \mathrm{d}x = \frac{2}{1+u^2}\mathrm{d}u$ |
-| 17 | $\displaystyle\int \frac{a\cos x+ b\sin x}{c\cos x + d \sin x} = A\int\mathrm{d}x+B\int\frac{\mathrm{d}(c\cos x+d \sin x)}{c \cos x + d \sin x}$ <br> (设 $a\cos x + b\sin x = A(c \cos x + d \sin x) + B(c \cos x + d \sin x)'$，由待定系数法求出A、B)$ |

source/_posts/2023-04-14-josh-adv-math-1-1.md

@@ -0,0 +1,285 @@
+---
+title: Josh's Summary — 工科数学分析<br>Part 1.1 基本微分与积分
+mathjax: true
+comments: true
+copyright: true
+toc:
+  enable: true
+  number: false
+  wrap: true
+  expand_all: true
+  max_depth: 4
+abbrlink: 1823c9d1
+date: 2023-04-14 16:25:59
+tags:
+  - 等价无穷小
+  - 导数公式
+  - Maclaurin 公式
+  - 曲率
+  - 积分公式
+  - 三角函数
+categories:
+  - - Josh 的学习笔记
+    - 微积分
+  - - 数学
+    - 微积分
+---
+
+$\newcommand\diff{\mathop{}\!\mathrm{d}}$
+
+# 1. 常见等价无穷小
+
+||||
+| :--: | :--: | :--: |
+| $\sin x \sim x$ | $\tan x \sim x$ | $\arcsin x \sim x$ |
+| $\arctan x \sim x$ | $\displaystyle\left( 1 - \cos x \right) \sim \dfrac{x^2}{2}$ | $\ln (1+x) \sim x$ |
+| $\left( e^x -1 \right) \sim x$ | $\left( a^x - 1 \right) \sim x \ln a$ | $\left( \left( 1+x \right) ^\alpha -1 \right) \sim \alpha x$  |
+
+<!-- more -->
+
+# 2. 导数基本公式
+
+| 一般函数 | 三角函数 | 反三角函数 | 双曲函数<br>(圆函数) |
+| :--: | :--: | :--: | :----: |
+| $\left( C \right)' = 0$ | $\left( \sin x \right)' = \cos x$ | $\displaystyle\left( \arcsin x \right)' = \dfrac{1}{\sqrt{1-x^2}}$ | $\displaystyle\left( \sinh x \right)' = \left(\dfrac{e^x-e^{-x}}{2}\right)' = \cosh x$ |
+| $\left( x^\alpha \right)'=\alpha x^{\alpha-1} \left( \alpha \in  \mathbb{Z} \right)$ | $\left( \cos x \right)' = -\sin x$ | $\displaystyle\left( \arccos x \right)' = -\dfrac{1}{\sqrt{1-x^2}}$ | $\displaystyle\left( \cosh x \right)' = \left( \dfrac{e^x+e^{-x}}{2} \right)'=\sinh x$ |
+| $\displaystyle\left( \log_ax\right)'=\dfrac{1}{x\ln a}$ | $\displaystyle\left( \tan x \right)' = \dfrac{1}{\cos^2 x}=\sec^2 x$ | $\displaystyle\left( \arctan x \right)' = \dfrac{1}{1+x^2}$ | $\displaystyle\left( \mathrm{arsinh} \, x \right)' = \left(\ln \left( x+\sqrt{x^2 + 1} \right) \right)' = \dfrac{1}{x^2+1}$ |
+| $( \ln$\| $x$ \|$)'=\displaystyle\dfrac{1}{x}$ | $\displaystyle\left( \cot x \right)' = -\dfrac{1}{\sin^2 x} = -\csc^2 x$ | $\displaystyle\left( \mathrm{arccot} \, x \right)' = -\dfrac{1}{1+x^2}$ | $\displaystyle\left( \mathrm{arcosh} \, x \right)' = \left( \ln \left( x + \sqrt{x^2 - 1} \right) \right)' = \dfrac{1}{\sqrt{x^2-1}}$
+|  | $\left( \sec x \right)' = \sec x \cdot \tan x$ |
+|  | $\left( \csc x \right)' = \csc x \cdot \cot x$ |
+
+- Newton-Leibniz 公式：
+
+  $$
+  \begin{aligned}
+  \left( u \cdot v \right)^{\left( n \right)} &= u^{\left( n \right)}v + nu^{\left( n-1 \right)}v' + \dfrac{n \left( n-1 \right)}{2!}u^{\left( n-2 \right)}v'' + \cdots + \dfrac{n\left( n-1 \right)\cdots\left(n-k+1\right)}{k!}u^{\left( n-k \right)}v^{(k)} + \cdots + uv^{\left( n \right)} \\
+  &= \sum_{k=0}^{n}C_n^k u^{\left( u-k \right)}v^k
+  \end{aligned}
+  $$
+
+  可以注意到 Newton-Leibniz 公式具有类似二项式定理展开的形式。
+
+# 3. 几个初等函数的 Maclaurin 公式
+
+## 3.1 Taylor 公式的一般形式
+
+$$\begin{aligned}
+f(x)&=f(x_0) + f'(x_0)(x-x_0) + \dfrac{f''(x_0)}{2!}(x-x_0)^2 + \cdots + \dfrac{f^{(n)}(x_0)}{n!}(x-x_0)^n + R_n
+\\
+&= \sum_{k=0}^n \dfrac{f^{(n)}(x_0)}{n!}(x - x_0)^n + R_n
+\end{aligned}$$
+
+其中
+
+$$
+R_n = \dfrac{f^{(n+1)}(\xi)}{(n+1)!}(x-x_0)^{n+1} \xlongequal{x\to x_0} o\left( \left( x - x_0 \right)^n \right), \quad \xi\,在\,x\,和\,x_0\,之间
+$$
+
+## 3.2 Maclaurin 公式的一般形式
+
+$$\begin{aligned}
+f(x) &= f(0)+f'(0)\,x + \dfrac{f''(0)}{2!}x^2 + \cdots + \dfrac{f^{(n)}(0)}{n!}x^n + R_n
+\\
+&= \sum_{k=0}^n \dfrac{f^{(n)}(0)}{n!}x^n + R_n
+\end{aligned}$$
+
+其中
+
+$$
+R_n = \dfrac{f^{(n+1)}(\xi)}{(n+1)!}x^{n+1} = \dfrac{f^{(n+1)}(\theta x)}{(n+1)!} x^{n+1} \xlongequal{x\to x_0} o\left( \left( x - x_0 \right)^n \right), \quad \xi\,在\,0\,和\,x\,之间, 0<\theta<1
+$$
+
+## 3.3 指数函数的 Maclaurin 展开式
+
+$$
+e^x=1+x+\dfrac{x^2}{2} + \dfrac{x^3}{3} + \cdots +\dfrac{x^n}{n!} + \dfrac{e^{\theta x}}{(n+1)!}, \quad 0<\theta<1
+$$
+
+## 3.4 正弦函数的 Maclaurin 展开式
+
+$$
+\sin x = x - \dfrac{x^3}{3!} + \dfrac{x^5}{5!} + \cdots + (-1)^{m-1}\dfrac{x^{2m-1}}{(2m-1)!} + R_{2m}
+$$
+
+其中
+
+$$
+R_{2m}=\dfrac{\sin\left( \theta x + \dfrac{(2m-1)\pi}{2}\right)}{(2m+1)!}x^{2m+1} = o(x^{2m}), \quad 0<\theta<1
+$$
+
+## 3.5 余弦函数的 Maclaurin 展开式
+
+$$
+\cos x = 1 - \dfrac{x^2}{2!} + \dfrac{x^4}{4!} + \cdots + (-1)^m\dfrac{x^{2m}}{2m!} + R_{2m+1}
+$$
+
+其中
+
+$$
+R_{2m+1}=\dfrac{\cos(\theta x + (m+1)\pi)}{(2m+2)!} x^{2m+2}=o(x^{2m+2}), \quad 0<\theta<1
+$$
+
+## 3.6 $\ln (1+x)$ 的 Maclaurin 展开式
+
+$$
+\ln(1+x) = x - \dfrac{x^2}{2} + \dfrac{x^3}{3} -\dfrac{x^4}{4} + \cdots + \dfrac{(-1)^{n-1}x^n}{n} + R_n(x)
+$$
+
+其中
+
+$$
+R_n(x) = \dfrac{(-1)^n}{(n+1)(1+\theta x)^{n+1}}x^{n+1} = o(x^n), \quad 0<\theta<1
+$$
+
+## 3.7 $(1+x)^\alpha$ 的 Maclaurin 展开式
+
+$$
+(1+x)^\alpha = 1 + \alpha x + \dfrac{\alpha(\alpha-1)}{2!}x^2 + \cdots + \dfrac{\alpha(\alpha - 1)\cdots (\alpha-n+1)}{n!}x^n + R_n(x)
+$$
+
+其中
+
+$$
+R_n(x) = \dfrac{\alpha(\alpha -1)\cdots(\alpha-n)}{(n+1)!}(1+\theta x)^{\alpha-n-1}x^{n+1} = o(x), \quad 0<\theta<1
+$$
+
+## 3.8 $\dfrac{1}{1-x}$ 的 Maclaurin 展开式
+
+$$
+\dfrac{1}{1-x} = 1+x+x^2+\cdots+x^n +o(x^n)
+$$
+
+# 4. 曲率及曲率半径
+
+| 曲率 | 曲率半径 |
+| :--: | :--: |
+| $K=\left\| \dfrac{\diff \alpha}{\diff s}\right\| \xlongequal{令\ \tan \alpha = y', 则\ \alpha = \arctan y'} \dfrac{\left\| \dfrac{y''}{1+y'^2}\diff x \right\|}{\sqrt{1+y'^2}}\diff x = \dfrac{\left\| y'' \right\|}{\left( 1 + y' \right)^{3/2}}$ | $R = \dfrac{1}{K}$ |
+
+# 5. 积分常用公式
+
+| 序 | 描述 | 公式 |
+| :--: | :----: | :----------------------------------------------:|
+| 1 | 定积分的定义 | $\displaystyle\int^1_0f(x)\diff x=I=\lim_{\lambda \to 0}\sum_{i=1}^{n}f(\xi_i)\Delta x_i$，其中 $\lambda = \max \left\{ \Delta x_1,\Delta x_2, \cdots, \Delta x_n \right\}$ |
+| 2 | 变上限积分 | $考察定积分\displaystyle\int^x_af(x)\diff x=\int^x_af(t)\diff t$，对 $\displaystyle\Phi(x)=\int_a^xf(t)\diff t$，有  $\displaystyle\Phi'(x)=\dfrac{\diff }{\diff x}\int_a^xf(t)\diff t = f(x)$ |
+| 3 | 变上限积分的扩展 | $\displaystyle\dfrac{\diff }{\diff x}\int^{\beta(x)}_{\alpha(x)}f'(t)\diff t = f\left[ \beta\left(x\right)\right]\beta'(x)-f[\alpha(x)]\alpha'(x)$ |
+| 4 | 上下限对称的积分 | 若 $f(x)$ 为偶函数，则 $\displaystyle\int_{-a}^af(x)\diff x=2\int_0^af(x)\diff x$；若 $f(x)$ 为奇函数，则 $\displaystyle\int_{-a}^af(x)\diff x=0$ |
+| 5 | 周期函数的积分 | $\displaystyle f(x+T)=f(x)\Rightarrow\int_0^{nT}f(x)\diff x=n\int^T_0f(x)\diff x$ |
+| 6 | 分部积分法 | $\displaystyle\int uv'\diff x=uv - \int u'v\diff x \Leftrightarrow \int u\diff v = uv - \int v\diff u$ |
+
+# 6. 基本积分表
+
+| 序 | 积分式 |
+| :--: | :-------------------------: |
+| 1 | $\displaystyle\int k\diff x=kx+C$，特别地 $\displaystyle\int\diff x=x+C$ |
+| 2 | $\displaystyle\int x^\mu\diff x=\dfrac{x^{\mu+1}}{\mu+1}+C(\mu\ne-1)$，特别地 $\displaystyle\int \dfrac{1}{x^2}\diff x=-\dfrac{1}{x}+C,\int\dfrac{1}{\sqrt{x}}\diff x=2\sqrt{x}+C$ |
+| 3 | $\displaystyle\int\dfrac{1}{x}\diff x=\ln \|x\|+C$ |
+| 4 | $\displaystyle\int\dfrac{1}{1+x^2}\diff x = \arctan x +C$ |
+| 5 | $\displaystyle\int\dfrac{1}{\sqrt{1-x^2}}\diff x=\arcsin x+C$ |
+| 6 | $\displaystyle\int\sin x\diff x = -\cos x+C, \int\cos x\diff x = \sin x +C$ |
+| 7 | $\displaystyle\int\dfrac{\diff x}{\cos^2x} = \int\sec^2x\diff x = \tan x +C, \int\dfrac{\diff x}{\sin^2x}=\int\csc^2x\diff x = -\cot x +C$ |
+| 8 |  $\displaystyle\int a^x\diff x = \dfrac{a^x}{\ln a} + C$，特别地 $\displaystyle\int e^x\diff x = e^x +C$
+| 9 | $\displaystyle\int \sinh x\diff x = \cosh x + C，\int \cosh x \diff x=\sinh x +C$ |
+
+# 7. 三角函数相关公式
+
+| 序 | 公式 |
+| :--: | :-------------------------: |
+| 1 | $\sin^2x+\cos^2x=1, \tan^2x+1=\sec^2x,\cot^2x+1=\csc^2x$ |
+| 2 | $\begin{aligned}&\cosh^2x-\sinh^2x=1,\\ &\sinh(x\pm y)=\sinh x\cdot\cosh y \pm \sinh y\cdot\cosh x, \cosh(x\pm y) = \cosh x\cdot\cosh y,\\ &\sinh2x = 2\sinh x\cdot \cosh x, \cosh 2x = \cosh^2x+\sinh^2x = 2\cosh^2x - 1\end{aligned}$ |
+| 3<br>积化和差公式 | $\sin\alpha\cos\beta=\dfrac12[\sin(\alpha+\beta)+\sin(\alpha-\beta)],\cos\alpha\sin\beta = \dfrac12[\sin(\alpha+\beta) - \sin(\alpha-\beta)]$ <br> $\cos\alpha\cos\beta = \dfrac12[cos(\alpha+\beta) + cos(\alpha - \beta)], \sin\alpha\sin\beta=-\dfrac12[\cos(\alpha+\beta) - \cos(\alpha-\beta)]$ |
+| 4<br>和差化积公式 | $\displaystyle\sin\alpha + \sin\beta = 2\sin\left(\dfrac{\alpha+\beta}{2}\right) \cos\left(\dfrac{\alpha-\beta}{2}\right), \sin\alpha - \sin\beta = 2\sin\left(\dfrac{\alpha-\beta}{2}\right) \cos\left(\dfrac{\alpha+\beta}{2}\right)$ <br> $\displaystyle\cos\alpha + \cos\beta = 2\cos\left(\dfrac{\alpha+\beta}{2}\right) \cos\left(\dfrac{\alpha-\beta}{2}\right), \cos\alpha - \cos\beta = -2\sin\left(\dfrac{\alpha+\beta}{2}\right) \sin\left(\dfrac{\alpha-\beta}{2}\right)$ |
+| 5<br>三角代换 | $\displaystyle\sqrt{a^2-x^2}\xrightarrow{x=a\sin t}a\cos t, \sqrt{a^2+x^2} \xrightarrow{x=a\tan t}a\sec t, \sqrt{x^2-a^2}\xrightarrow{x=a\sec t}a\tan t \left( a>0,t\in\left( -\dfrac{\pi}{2} , \dfrac{\pi}{2} \right) \right)$ |
+| 6 | $\displaystyle\int\sqrt{a^2-x^2}\diff x = \dfrac{a^2}{2}\arcsin\dfrac xa + \dfrac12x\sqrt{a^2-x^2} + C(a>0)$ |
+| 7 | $\displaystyle\int\sqrt{x^2+a^2}\diff x = \dfrac x2\sqrt{x^2+a^2} + \dfrac{a^2}2\ln \left\| x+\sqrt{x^2+a^2} \right\|+C(a>0)$ |
+| 8 | $\displaystyle\int\dfrac{1}{\sqrt{x^2+a^2}}\diff x = \ln\left\| x+ \sqrt{x^2+a^2} \right\| +C$, $\displaystyle\int\dfrac{1}{\sqrt{x^2-a^2}}\diff x = \ln\left\| x + \sqrt{x^2-a^2}\right\|(a>0)$<br>特别地，$\displaystyle\int\dfrac{1}{\sqrt{x^2+1}}\diff x=\ln\left(x+\sqrt{x^2+1}\right) = \mathrm{arsinh}\, x$, $\displaystyle\int\dfrac{1}{\sqrt{x^2-1}}\diff x = \ln\left( x+\sqrt{x^2-1}\right) = \mathrm{arcosh}\, x$ |
+| 9 | $\displaystyle\int\sec x\diff x = \int\dfrac{1}{\cos x}\diff x = \int\dfrac{\sec^2x+\sec x \tan x}{\sec x + \tan x}\diff x = \int\dfrac{1}{\sec x + \tan x}\diff (\sec x + \tan x) = \ln \left\| \sec x + \tan x \right\| + C$ |
+| 10 | $\displaystyle\int\csc x\diff x = \int\dfrac{1}{\sin x}\diff x = \ln\left\| \tan\dfrac x2 \right\| + C = \dfrac 12 \ln \left\| \dfrac{1-\cos x}{1 + \cos x}\right\| + C$ |
+| 11 | $\displaystyle\int\tan x\diff x = \int\dfrac{\sin x}{\cos x}\diff x = -\int\dfrac{1}{\cos x}\diff (\cos x) = -\ln \left\| \cos x\right\| +C$ |
+| 12 | $\displaystyle\int\dfrac{1}{\tan x}\diff x = \int\dfrac{\cos x}{\sin x}\diff x = \int\dfrac{1}{\sin x}\diff (\sin x) = \ln\left\| \sin x \right\| + C$ |
+| 13 | $\displaystyle\int\arctan x\diff x = x\arctan x - \ln \left\| 1+x^2 \right\| +C$ |
+| 14<br>Wallis 公式 | $\displaystyle\int^{\pi/2}_{0}\sin^nx\diff x = \int^{\pi/2}_{0}\cos^nx\diff x = \begin{cases} \displaystyle\dfrac{n-1}{n}\cdot\dfrac{n-3}{n-2}\cdot\cdots\cdot\dfrac12\cdot\dfrac\pi2,&n为偶数 \\ \displaystyle\dfrac{n-1}{n}\cdot\dfrac{n-3}{n-2}\cdot\cdots\cdot\dfrac23\cdot1,&n为奇数 \end{cases} \\ = \begin{cases} \displaystyle\dfrac{(n-1)!!}{n!!}\cdot\dfrac\pi2,&n为偶数 \\ \displaystyle\dfrac{(n-1)!!}{n!!},&n为奇数 \end{cases}$ |
+| 15 | $\displaystyle\int_0^\pi\sin^nx\diff x = 2\int^{\pi/2}_0\sin^nx\diff x,\int_0^\pi\cos^nx\diff x = \begin{cases} 2\displaystyle\int^{\pi/2}_0\cos^nx\diff x,&n为偶数 \\ 0,&n为奇数 \end{cases}$ <br> $\displaystyle\int_0^{2\pi}\sin^nx = \int_0^{2\pi}\cos^nx\diff x = \begin{cases} 4\displaystyle\int^{\pi/2}_0\sin^nx\diff x,&n为偶数 \\ 0,&n为奇数 \end{cases}$ <br> $\displaystyle\int_0^{\pi/2}f(\sin x)\diff x = \int_0^{\pi/2}f(\cos x)\diff x$, $\displaystyle\int_0^\pi f(\sin x)\diff x \ne \int_0^\pi f(\cos x)\diff x$ <br> $\displaystyle\int_0^\pi\pi f(\sin x)\diff x = \dfrac\pi2\int_0^\pi f(\sin x)\diff x = \pi\int_0^{\pi/2} f(\sin x)\diff x$ |
+| 16<br>万能公式 | 令\ $u=\displaystyle\tan\dfrac x2,x=2\arctan u$，则\ $\displaystyle\sin x = \dfrac{2u}{1+u^2},\cos x=\dfrac{1-u^2}{1+u^2}, \diff x = \dfrac{2}{1+u^2}\diff u$ |
+| 17 | $\displaystyle\int \dfrac{a\cos x+ b\sin x}{c\cos x + d \sin x} = A\int\diff x+B\int\dfrac{\diff (c\cos x+d \sin x)}{c \cos x + d \sin x}$ <br> (设 $a\cos x + b\sin x = A(c \cos x + d \sin x) + B(c \cos x + d \sin x)'$，由待定系数法求出 $A, B$) |
+
+# 8. 定积分的几何应用
+
+## 8.1 平面
+
+### 8.1.1 直角坐标
+
+1. 直角梯形的面积
+
+$$\begin{equation*}
+A = \int_a^b [f(x_1) - f(x_2)] \diff x
+\end{equation*}$$
+
+2. 曲边梯形的面积：对于参数方程 $\begin{cases}
+     x = \varphi(x) \\ y = {\psi(x)}
+   \end{cases}$，曲边梯形的面积（变换上下限）
+
+   $$\begin{equation*}
+   A = \int_a^b y\diff x = \int_{t_1}^{t_2} \psi(t) \diff \varphi(t) = \int_{t_1}^{t_2} \psi(t)\varphi'(t) \diff t
+   \end{equation*}$$
+
+3. 椭圆的面积：椭圆 $\begin{cases}
+     x = a \cos t \\ y = b \sin t
+   \end{cases}$，即 $\dfrac{x^2}{a^2} + \dfrac{y^2}{b^2} = 1$ 的面积
+
+   $$\begin{equation*}
+     A = 4 \int_0^a y \diff x = 4ab \int_0^{\pi/2} (1-\cos t)^2 \diff t = 3\pi a^2
+   \end{equation*}$$
+
+4. 摆线的面积：摆线 $\begin{cases}
+     x = a (t - \sin t) \\ y = a (1 - \cos t)
+   \end{cases}$ 的一拱与 $x$ 轴所围图形的面积
+
+   $$\begin{equation*}
+     A = \int_0^2\pi a y \diff x = a^2 \int_0^{2\pi} (1 - \cos t)^2 \diff t = 3\pi a^2
+   \end{equation*}$$
+
+### 8.1.2 极坐标
+
+1. 扇形的面积：由 $\rho  = \rho(\theta)$ 及射线 $\theta = \alpha$ 和 $\theta = \beta$ 所围曲边扇形的面积
+
+   $$\begin{equation*}
+      A = \int_\alpha^\beta \frac{1}{2} \rho^2(\theta) \diff \theta = \frac{1}{2} \int_\alpha^\beta \rho^2 \diff \theta
+   \end{equation*}$$
+
+2. 阿基米德螺线：阿基米德螺线 $\rho = a \theta (a > 0)$ 对应 $\theta$ 从 $0$ 变换到 $2\pi$ 所围图形的面积
+
+   $$\begin{equation*}
+      A = \int_0^{2\pi} \frac{1}{2} (a\theta)^2\diff \theta = \frac{4}{3} \pi^3 a^2
+   \end{equation*}$$
+
+3. 双纽线：双纽线 $\rho^2 = a^2 \cos 2\theta$ 所围平面图形的面积
+
+   $$\begin{equation*}
+      A = 4 \int_0^{\pi/4} a^2 \cos 2\theta \diff \theta = a^2
+   \end{equation*}$$
+
+4. 心形线：心形线 $\rho = a(a+\cos\theta), a>0$ 所围平面图形的面积
+
+   $$\begin{equation*}
+      A = 2 \cdot \frac{1}{2} \int_0^\pi a^2 (1+ \cos \theta)^2 \diff \theta = \frac{3}{2}\pi a^2
+   \end{equation*}$$
+
+5. 曲线的弧长
+
+   | 弧微分 | 直角坐标 | 参数方程 | 极坐标 |
+   | :--: | :--: | :--: | :--: |
+   | $\diff s = \sqrt{(\diff x)^2 + (\diff y)^2}$ | $s = \displaystyle\int_a^b \sqrt{1 + y'^2}\diff x$ | $s = \displaystyle\int_a^b \sqrt{\varphi'^2(t) + \psi'^2(t)}\diff t$ | $s = \displaystyle\int_\alpha^\beta \sqrt{\rho^2{\theta} + \rho'^2(\theta)} \diff\theta$
+
+## 8.2 空间
+
+1. 已知平行截面面积求立体体积：令 $A(x)$ 为截面面积，则立体体积
+
+   $$\begin{equation*}
+      V = \int_a^b A(x) \diff x
+   \end{equation*}$$
+
+2. 旋转体体积
+
+   | 薄片法 | 柱壳法 |
+   | :--: | :--: |
+   | $V = \displaystyle\int_a^b \pi f^2(x) \diff x$ | $V = \displaystyle\int_a^b 2\pi x \left| f(x) \right|\diff x$ |

source/_posts/2025-07-02-gaussian-dist-pdf-derivation.md

@@ -0,0 +1,279 @@
+---
+title: 信息论角度推导高斯分布的概率密度函数
+mathjax: true
+comments: true
+copyright: true
+toc:
+  enable: true
+  number: true
+  wrap: true
+  expand_all: true
+  max_depth: 4
+abbrlink: '741621e3'
+date: 2025-07-02 09:20:07
+tags:
+  - 高斯分布
+  - 信息论
+categories:
+  - [数学, 概率论]
+---
+
+$\newcommand\diff{\mathop{}\!\mathrm{d}}$
+
+# 背景与目标
+
+在统计建模中，我们常常面临这样一个问题：**已知某个连续随机变量的期望和方差，如何合理地确定它的概率分布？** 最大熵原理（Maximum Entropy Principle）提供了一个具有信息论意义的答案：
+
+> 在已知约束（如期望与方差）下，熵最大的概率密度函数是最“中立”或最“无偏”的选择。
+
+熵在此表示不确定性或信息量。连续型随机变量的微分熵定义为：
+
+$$
+\begin{equation*}
+H[p] = -\int_{-\infty}^{\infty} p(x) \ln p(x) \, \diff x
+\end{equation*}
+$$
+
+我们的目标是：**在已知期望和方差的约束下，最大化熵 $H[p]$，求出最优概率密度函数 $p(x)$**。
+
+<!-- more -->
+
+# 约束条件
+
+我们假设随机变量 $X$ 满足：
+
+- 归一化约束：
+
+  $$
+  \begin{equation*}
+  \int_{-\infty}^{\infty} p(x) \, \diff x = 1
+  \end{equation*}
+  $$
+
+- 期望约束：
+
+  $$
+  \begin{equation*}
+  \int_{-\infty}^{\infty} x \, p(x) \, \diff x = \mu
+  \end{equation*}
+  $$
+
+- 方差约束：
+
+  $$
+  \begin{equation*}
+  \int_{-\infty}^{\infty} (x - \mu)^2 \, p(x) \, \diff x = \sigma^2
+  \end{equation*}
+  $$
+
+# 构造拉格朗日泛函
+
+引入拉格朗日乘子 $\lambda_0, \lambda_1, \lambda_2$，构造带约束的目标泛函：
+
+$$
+\begin{equation*}
+\mathcal{L}[p] = -\int p(x) \ln p(x) \diff x
++ \lambda_0 \left( \int p(x) \diff x - 1 \right)
++ \lambda_1 \left( \int x p(x) \diff x - \mu \right)
++ \lambda_2 \left( \int (x - \mu)^2 p(x) \diff x - \sigma^2 \right)
+\end{equation*}
+$$
+
+将所有积分项合并为：
+
+$$
+\begin{equation*}
+\mathcal{L}[p] = \int \left[
+- p(x) \ln p(x)
++ \lambda_0 p(x)
++ \lambda_1 x p(x)
++ \lambda_2 (x - \mu)^2 p(x)
+\right] \diff x
+\end{equation*}
+$$
+
+# 变分法求极值
+
+我们考虑一个扰动：
+
+$$
+\begin{equation*}
+p(x) \rightarrow p(x) + \epsilon \eta(x)
+\end{equation*}
+$$
+
+其中 $\epsilon$ 是足够小的实数，$\eta(x)$ 是任意光滑、在积分区间内有界且满足边界条件的函数（可理解为“任意小的扰动”）。为了求出 $\dfrac{\delta \mathcal{L}}{\delta p(x)}$，我们代入扰动项进行展开：
+
+$$
+\begin{equation*}
+\mathcal{L}[p + \epsilon \eta]
+= -\int (p(x) + \epsilon \eta(x)) \ln (p(x) + \epsilon \eta(x)) \, \diff x + \text{其余约束项}
+\end{equation*}
+$$
+
+> 注意：这里只展开了主熵项 $-\displaystyle\int p \ln p$，约束项形式类似，将在后面统一处理。
+
+---
+
+对主项使用泰勒展开（忽略高阶小量）：
+
+> **泰勒展开回顾：**
+>
+> 设 $f(p) = p \ln p$，我们对 $p(x) + \epsilon \eta(x)$ 展开，有：
+> $$
+> f(p + \epsilon \eta) = p \ln p + \epsilon \eta (1 + \ln p) + o(\epsilon)
+> $$
+> 其中：
+> $$
+> \frac{\diff }{\diff p}(p \ln p) = \ln p + 1
+> $$
+
+将其代入：
+
+$$
+\begin{align*}
+\mathcal{L}[p + \epsilon \eta]
+&= -\int \left[ p(x) \ln p(x) + \epsilon \eta(x)(1 + \ln p(x)) \right] \diff x + o(\epsilon) + \text{约束项} \\
+&= \mathcal{L}[p] - \epsilon \int \eta(x) (1 + \ln p(x)) \diff x + o(\epsilon) + \text{约束项}
+\end{align*}
+$$
+
+对约束项也类似展开：
+
+- 对归一化项：
+
+  $$
+  \begin{equation*}
+  \lambda_0 \left( \int (p + \epsilon \eta) \diff x - 1 \right)
+  = \lambda_0 \left( \int p \, \diff x - 1 + \epsilon \int \eta \, \diff x \right)
+  \end{equation*}
+  $$
+
+- 对期望项：
+
+  $$
+  \begin{equation*}
+  \lambda_1 \left( \int x (p + \epsilon \eta) \diff x - \mu \right)
+  = \lambda_1 \left( \int x p \, \diff x - \mu + \epsilon \int x \eta \, \diff x \right)
+  \end{equation*}
+  $$
+
+- 对方差项：
+
+  $$
+  \begin{equation*}
+  \lambda_2 \left( \int (x - \mu)^2 (p + \epsilon \eta) \diff x - \sigma^2 \right)
+  = \lambda_2 \left( \int (x - \mu)^2 p \, \diff x - \sigma^2 + \epsilon \int (x - \mu)^2 \eta(x) \diff x \right)
+  \end{equation*}
+  $$
+
+---
+
+将所有一阶 $\epsilon$ 项收集起来：
+
+$$
+\begin{equation*}
+\mathcal{L}[p + \epsilon \eta] = \mathcal{L}[p]
++ \epsilon \int \eta(x) \left[ - (1 + \ln p(x)) + \lambda_0 + \lambda_1 x + \lambda_2 (x - \mu)^2 \right] \diff x + o(\epsilon)
+\end{equation*}
+$$
+
+对 $\epsilon$ 求导并令其在 $\epsilon = 0$ 处为零：
+
+$$
+\begin{equation*}
+\left. \frac{\diff}{\diff \epsilon} \mathcal{L}[p + \epsilon \eta] \right|_{\epsilon=0}
+= \int \eta(x) \left[ -\ln p(x) - 1 + \lambda_0 + \lambda_1 x + \lambda_2 (x - \mu)^2 \right] \diff x = 0
+\end{equation*}
+$$
+
+根据变分法基本引理，这个积分当且仅当括号内为 $0$ 时对所有扰动函数 $\eta(x)$ 为零：
+
+> 变分法基本引理：
+>
+> 若函数 $F(x)$ 满足
+> $$
+> \int_{a}^{b} F(x) \cdot \eta(x) dx = 0
+> $$
+> 对所有合适扰动 $\eta(x)$ 成立，则必有
+> $$
+> F(x) = 0 \quad \forall x \in (a, b)
+> $$
+>
+> 这是因为若 $F(x_0) \ne 0$，则可构造在 $x_0$ 附近支持的 $\eta(x)$ 使得积分不为零，导致矛盾。
+
+$$
+\begin{equation*}
+\frac{\delta \mathcal{L}}{\delta p(x)} = -\ln p(x) - 1 + \lambda_0 + \lambda_1 x + \lambda_2 (x - \mu)^2 = 0
+\end{equation*}
+$$
+
+也即求得了变分导数。移项得：
+
+$$
+\begin{equation*}
+\ln p(x) = \lambda_0' + \lambda_1 x + \lambda_2 (x - \mu)^2
+\end{equation*}
+$$
+
+两边取指数：
+
+$$
+\begin{equation*}
+p(x) = \exp\left( \lambda_0' + \lambda_1 x + \lambda_2 (x - \mu)^2 \right)
+\end{equation*}
+$$
+
+要满足可积性，需 $\lambda_2 < 0$。设
+
+$$
+\begin{equation*}
+p(x) = A \cdot \exp\left( -\alpha (x - \mu)^2 \right)
+\end{equation*}
+$$
+
+# 归一化与确定参数
+
+根据归一化条件：
+
+$$
+\begin{equation*}
+A \cdot \int_{-\infty}^{\infty} \exp\left( -\alpha (x - \mu)^2 \right) \diff x = 1
+\end{equation*}
+$$
+
+已知高斯积分结果：
+
+$$
+\begin{equation*}
+\int_{-\infty}^{\infty} \exp\left( -\alpha (x - \mu)^2 \right) \diff x = \sqrt{\frac{\pi}{\alpha}}
+\end{equation*}
+$$
+
+得到：
+
+$$
+\begin{equation*}
+A = \sqrt{\frac{\alpha}{\pi}}
+\end{equation*}
+$$
+
+若令 $\alpha = \dfrac{1}{2\sigma^2}$，则最终得：
+
+$$
+\begin{equation*}
+p(x) = \frac{1}{\sqrt{2\pi\sigma^2}} \exp\left( -\frac{(x - \mu)^2}{2\sigma^2} \right)
+\end{equation*}
+$$
+
+# 总结
+
+从最大熵原理出发，在已知均值与方差的约束下，最“无偏”或熵最大的概率密度函数是：
+
+$$
+\begin{equation*}
+p(x) = \frac{1}{\sqrt{2\pi\sigma^2}} \exp\left( -\frac{(x - \mu)^2}{2\sigma^2} \right)
+\end{equation*}
+$$
+
+这正是**高斯分布**的标准形式，说明它不仅数学上方便，**信息论上也是最合理的选择**。