Mathematical Foundation series (II) ---- partial derivatives, directional derivative, gradient, calculus

A partial derivatives

For unary function y = f (x) exists only y varies with x, but binary function z = f (x, y) exists z change ratio x changes, with y varying the rate of change, with x ﹑ y simultaneously changes in the rate of change. As shown below

  

1, defined partial derivative

Provided function $ z = f (x, y ) $ at the point (X 0 , Y 0 ) of a neighborhood is defined, set Y = Y 0 , unary functions $ f (x_ {0}, y_ {0}) $ at the point X = X 0 at a derivable, i.e. the limit $ \ lim \ limits _ {\ Delta x \ rightarrow 0} \ frac {f (x_ {0} + \ Delta x, y_ {0}) - f (x_ { 0}, y_ {0}) } {\ Delta x} = A $.

A function is called $ z = f (x, y ) $ at the point (x 0 , Y 0 ) of the argument x on partial derivatives . Referred to as: $ f_ {x} (x_ {0}, y_ {0}) $, $ \ frac {\ partial z} {\ partial x} \ left | _ {\ begin {smallmatrix} x = {{x} . _ {0}} \\ y = {{y} _ {0}} \ end {smallmatrix}} \ right $, $ \ frac {\ partial f} {\ partial x} \ left | _ {\ begin { smallmatrix} x = {{x} _ {0}} \\ y = {{y} _ {0}} \ end {smallmatrix}} \ right $ or $ z_ {x} \ left | . _ {\ begin { smallmatrix} x = {{x} _ {0}} \\ y = {{y} _ {0}} \ end {smallmatrix}} \ right. $

2, geometric meaning :

Partial derivative $ f_ {x} (x_ { 0}, y_ {0}) $ is a curved surface is a plane $ y = y_ {0} $ are intercepted curve at point M 0 tangent at M 0 T x x-axis the slope , the partial derivative $ f_ {y} (x_ { 0}, y_ {0}) $ is a curved surface is planarized $ x = x_ {0} $ are intercepted curve at point M 0 tangent at M 0 Ty of y-axis inclination . As shown below

  

3, example

Request $ f (x, y) = x ^ {2} + 3xy + y ^ {2} $ point (1, 2) partial derivatives.

  

Second, the directional derivative

1 Introduction

Point in the domain of the function of the derivative guide obtained on demand a certain direction. Direction generally function binary function and ternary derivative, directional derivative directional derivative can be divided in the linear direction and the direction of the curve. Now suppose that, as shown below, there are two flames along the x, y-axis spread, run to survive ant asked what direction?

  

You can easily expect to run along the diagonal of the rectangle. There are a function of $ z = f (x, y) $

  

It can be derived from $ \ left | PP '\ right | = \ rho = \ sqrt {(\ Delta x ^ {2}) + (\ Delta y ^ {2})} $, may then be derived function value increases an amount of $ \ Delta z = f (x + \ Delta x, y + \ Delta y) -f (x, y) $.

If the delta function, the presence ratio of these two distances, this is called in along the direction L of the derivative at point P , is expressed by formulas $ \ frac {\ partial f} {\ partial l} = \ lim \ limits _ {\ rho \ rightarrow 0 } \ frac {f (x + \ Delta x, y + \ Delta y) -f (x, y)} {\ rho} $

In particular, the function $ f (x, y) $ in the positive X axis $ \ vec {e_ {1} } $ = {1,0}, Y -axis positive $ \ vec {e_ {2} } $ = { 0,1} directional derivatives are $ f_ {x}, f_ { y} $, negative direction derivative $ -f_ {x}, - f_ {y} $

2, Theorem

如果函数$z=f(x,y)$在点$P(x,y)$是可微分的,那么在该点沿任意方向L的方向导数都存在,公式表达为$\frac{\partial f}{\partial l}=\frac{\partial f}{\partial x}\cos \varphi +\frac{\partial f}{\partial y}\sin \varphi $,$\varphi $为X轴到L的角度。

  

3、例题

求函数$z=xe^{2y}$在点$P(1,0)$处沿从点$P(1,0)$到点$Q(2,-1)$的方向的方向导数。

示意图如下:

  

求解过程如下:

  

三、梯度

1、梯度

函数$z=f(x,y)$在平面域内具有连续的一阶偏导数,对于其中每一个点$P(x,y)$都有向量$\frac{\partial f}{\partial x}\vec{i}+\frac{\partial f}{\partial y}\vec{j}$,则其称为函数在点P的梯度。梯度的本意是一个向量(矢量),表示某一函数在该点处的方向导数沿着该方向取得最大值,即函数在该点处沿着该方向(此梯度的方向)变化最快,变化率最大(为该梯度的模)。用公式表达来就是:

$gradf(x,y)=\frac{\partial f}{\partial x}\vec{i}+\frac{\partial f}{\partial y}\vec{j}$。

设$\vec{e}=\cos \varphi \vec{i}+\sin \varphi \vec{j}$是方向L上的单位向量

由方向导数公式可知:$\frac{\partial f}{\partial l}=\frac{\partial f}{\partial x}\cos \varphi +\frac{\partial f}{\partial y}\sin \varphi=\left \{ \frac{\partial f}{\partial x}, \frac{\partial f}{\partial y}\right \}\cdot \left \{ \cos \varphi ,\sin \varphi  \right \}=gradf(x,y)\cdot \vec{e}=\left | gradf(x,y) \right |\cos\theta $。

其中$\theta =(gradf(x,y),\vec{e})$。

2、结论

只有$\cos (gradf(x,y),\vec{e})=1$,$\frac{\partial f}{\partial l}$才有最大值

函数在某点的梯度是一个向量,它的方向与方向导数最大值取得的方向一致,从而而它的模正好是最大的方向导数,也就是方向导数的最大值。

  

3、例题

设$u=xyz+x^{2}+5$,求grad$u$,并求在点$M(0,1,-1)$处方向导数的最大(小)值。

  

4、小结

(1)、方向导数的概念,注意方向导数与一般所说偏导数的区别
(2)、注意梯度其实是一个向量。
(3)、方向导数与梯度的关系,梯度的方向就是函数f(x,y)在这点增长最快的方向,梯度的模就是方向导数的最大值。

四、微积分

1、介绍

微积分诞生于17世纪,主要帮助人们解决各种速度,面积等实际问题,如下图所示,怎么才能求得曲线的面积呢?

  

首先对于一个矩形来说,我们可以轻松求得其面积,那能不能用矩形代替曲线形状呢?如果能行的话,那应该用多少个矩形来代替曲线呢?

   

在ab之间插入若干个点,这样就得到了n个小区间,这样的话每一个小矩形的面积为:$A_{i}=f(\xi _{i})\Delta x_{i}$,这样的话对每个小矩形的面积求和的话就可以近似得到曲线的面积:$A\approx \sum\limits_{i=1}^{n}f(\xi _{i})\Delta x_{i}$

   

当分割无限加细,每个小区间的最大长度为$\lambda $,此时$\lambda \rightarrow 0$。由此可得曲线的面积为:$A=\lim\limits_{\lambda  \rightarrow 0 }\sum\limits_{i=1}^{n}f(\xi _{i})\Delta x_{i}$

从求和角度来看,我们需要尽可能的将每一个矩形的底边无穷小,而莱布尼兹为了体现求和的感觉,给S拉长了,简写成$\int f(x)dx$

  

2、微分:

由于无穷小的概念,dx,dy都叫做微分。所谓微积分就是把这些微分积起来。

  

微分是什么?其实很简单,用两个式子就可以很简单的描述了:$\lim\limits_{\Delta x \rightarrow 0 }dy=0,\lim\limits_{\Delta x \rightarrow 0 }dx=0$

  

3、定积分

当|Δx|—>0时,总和S总是趋于确定的极限I,则称极限I为函数f(x)在曲线[a,b]上的定积分

  

其中,积分值和被积函数与积分曲线有关,与积分变量字母无关。

  

当函数f(x)在区间[a,b]上的定积分存在时,称f(x)在区间[a,b]上可积

4、定积分几何含义

面积的正负值:$\begin{matrix}f(x)>0, & \int_{a}^{b}f(x)dx=A\\ f(x)<0, & \int_{a}^{b}f(x)dx=-A\end{matrix}$

也就是说如果定积分的值为正值,那它就表示曲边梯形的面积,如果求出来是负值的话,那它就表示曲边梯形的面积的负值

  

5、定积分性质:

1、$\int_{a}^{b}[f(x)\pm g(x)]dx=\int_{a}^{b}f(x)dx\pm \int_{a}^{b}g(x)dx.$

2、$\int_{a}^{b}kf(x)dx=k\int_{a}^{b}f(x)dx$,k为常数。

3、假设a<c<b,则$\int_{a}^{b}f(x)dx=\int_{a}^{c}f(x)dx+\int_{c}^{b}f(x)dx$。

4、如果在区间[a,b]上$f(x)\geqslant 0$,则$\int_{a}^{b}f(x)dx\geqslant 0.(a<b)$。

Guess you like

Origin www.cnblogs.com/xiaoyh/p/12014108.html