cauchy problem of 1st order PDE from Partial Differential Equations

pure math

加一个一个Episodes

pde进可攻退可守

pure math

f : R → R , y = f ( x ) , d y = f ′ ( x ) d x f:\mathbb{R}\rightarrow\mathbb{R},y=f(x),dy=f'(x)dx f:R→R,y=f(x),dy=f′(x)dx

f ′ ( x ) ≠ 0 ⇒ x = f − 1 ( y ) f'(x)\neq0 \Rightarrow x=f^{-1}(y) f′(x)​=0⇒x=f−1(y)

A : R n → R n , y = A x , d y = A d x . A: \mathbb{R}^{n} \rightarrow \mathbb{R}^{n}, y=Ax,dy=Adx. A:Rn→Rn,y=Ax,dy=Adx.

d e t ( A ) ≠ 0 ⇒ x = A − 1 y det(A)\neq 0 \Rightarrow x=A^{-1}y det(A)​=0⇒x=A−1y

In general, you can not explicitly solve for the inverse!

f : R n → R n , y = f ( x ) , d y = D f d x . f:\mathbb{R}^n \rightarrow \mathbb{R}^n, y=f(x),dy=Dfdx. f:Rn→Rn,y=f(x),dy=Dfdx.

D f = [ δ ( y 1 , y 2 , y 3 , . . . y n ) ∂ ( x 1 , x 2 , x 3 , . . . , x n ) ] Df = [\frac{\delta(y_1,y_2,y_3,...y_n)}{\partial (x_1,x_2,x_3, ..., x_n)}] Df=[∂(x1​,x2​,x3​,...,xn​)δ(y1​,y2​,y3​,...yn​)​]

Jacobian matrix

d e t ( D f ) ≠ 0 ⇒ x = f − 1 ( y ) det(Df) \neq 0 \Rightarrow x=f^{-1}(y) det(Df)​=0⇒x=f−1(y)

在连续的函数中

一个点大于0

implies

一段大于0

theorem come

theorem go

but example last forever
这个代码的符号是在键盘的左上角的符号
三行五行的证明一定要会

但是两页的证明不要看了

一定要会用

用久了

一定会证明

t h e o r e m c o m e , t h e o r e m g o , b u t e x a m p l e l a s t f o r e v e r theorem come, theorem go, but example last forever theoremcome,theoremgo,butexamplelastforever

jacobian matrix

d x = x u d u + x v d v , d y = y u d u + y v d v dx=x_u du+x_v dv, dy = y_u du+y_v dv dx=xu​du+xv​dv,dy=yu​du+yv​dv

顺便吹爆这个math formula的插件,真香
cauchy problem of 1st order PDE from Partial Differential Equations

u x + 3 y 2 3 u y = 2 , u ( x , 1 ) = 1 + x u_x+3y^\frac{2}{3}u_y=2, u(x,1)=1+x ux​+3y32​uy​=2,u(x,1)=1+x

d x d t = 1 , d y d t = 3 y 2 3 , d u d t = 2 \frac{dx}{dt}=1, \frac{dy}{dt}=3y^\frac{2}{3},\frac{du}{dt}=2 dtdx​=1,dtdy​=3y32​,dtdu​=2

( x , y , u ) ∣ t = 0 = ( x 0 , y 0 , u 0 ) = ( s , 1 , 1 + s ) (x,y,u)|_{t=0}=(x_0,y_0,u_0)=(s,1,1+s) (x,y,u)∣t=0​=(x0​,y0​,u0​)=(s,1,1+s)

x = t + s , y = ( t + 1 ) 3 , u = 2 t + 1 + s x=t+s,y=(t+1)^3, u = 2t+1+s x=t+s,y=(t+1)3,u=2t+1+s

∂ ( x , y ) ∂ ( t , s ) \frac{\partial (x,y)}{\partial (t,s)} ∂(t,s)∂(x,y)​
cauchy problem of 1st order PDE from Partial Differential Equations

D ( q ) = [ p 1 + p 2 + 2 p 3 c o s q 2 p 2 + p 3 c o s q 2 p 2 + p 3 c o s q 2 p 2 ] D(q) = \begin{bmatrix} p_1+p_2+2p_3cosq_{2} & p_2+p_3cosq_2\\ p_2+p_3cosq_2 & p_2 \end{bmatrix} D(q)=[p1​+p2​+2p3​cosq2​p2​+p3​cosq2​​p2​+p3​cosq2​p2​​]
cauchy problem of 1st order PDE from Partial Differential Equations

C ( q , q ˙ ) = [ − p 3 q ˙ 2 s i n q 2 − p 3 ( q ˙ 1 + q ˙ 2 ) s i n q 2 p 3 q ˙ 1 s i n q 2 0 ] C(q,\dot q) = \begin{bmatrix} -p_3\dot q_2sinq_2 &-p_3(\dot q_1+\dot q_2)sinq_2 \\ p_3\dot q_1sinq_2 & 0 \end{bmatrix} C(q,q˙​)=[−p3​q˙​2​sinq2​p3​q˙​1​sinq2​​−p3​(q˙​1​+q˙​2​)sinq2​0​]

= ∣ x t x s y t y s ∣ = ∣ 1 1 3 ( t + 1 ) 2 0 ∣ = − 3 ( t + 1 ) 2 ≠ 0 \frac{}{}=\begin{vmatrix} x_t & x_s \\ y_t & y_s\end{vmatrix} = \begin{vmatrix} 1 & 1 \\ 3(t+1)^2 & 0\end{vmatrix} = -3(t+1)^2 \neq 0 ​=∣∣∣∣​xt​yt​​xs​ys​​∣∣∣∣​=∣∣∣∣​13(t+1)2​10​∣∣∣∣​=−3(t+1)2​=0

cauchy problem of 1st order PDE from Partial Differential Equations

t = y 1 3 − 1 , s = x + 1 − y 1 3 t=y^{\frac{1}{3}}-1, s=x+1-y^{\frac{1}{3}} t=y31​−1,s=x+1−y31​

eliminate s and t

u ( x , y ) = 2 t + 1 + s = x + y 1 3 u(x,y)=2t+1+s=x+y^{\frac{1}{3}} u(x,y)=2t+1+s=x+y31​

t ≠ − 1 t\neq -1 t​=−1

t = − 1 ⇒ y = 0 i s a s i n g u a l a r p o i n t o f D . E . t = -1 \Rightarrow y = 0 is a singualar point of D.E. t=−1⇒y=0isasingualarpointofD.E.

Dimensional Analysis: (from equation!)

[ u ] [ x ] = [ y ] 2 3 [ u ] [ y ] ⇒ [ x ] = [ y ] 1 3 \frac{[u]}{[x]}=[y]^{\frac{2}{3}}\frac{[u]}{[y]} \Rightarrow [x]=[y]^\frac{1}{3} [x][u]​=[y]32​[y][u]​⇒[x]=[y]31​

x → λ 2 x , y → λ y x \rightarrow \lambda^2 x, y \rightarrow \lambda y x→λ2x,y→λy

u ( u( u(

$$$$

$$$$

$$$$

$$$$

$$$$

$$$$

$$$$

$$$$

$$$$

$$$$

$$$$

上一篇:zookeeper源码分析之三客户端发送请求流程


下一篇:解密jQuery事件核心 - 委托设计(二)