Skip to main content

Linear First Order Differential Equations 線性一階微分方程式

線性微分方程式形式如下:

y+P(x)y=Q(x)y' + P(x)y = Q(x)

Q(x)=0Q(x) = 0 則稱為線性齊次微分方程式,通解:

y=CeP(x)dxy = C e^{-\int P(x) \, dx}

Q(x)0Q(x) \neq 0 則稱為線性非齊次微分方程式,通解:

y=eP(x)dx[Q(x)eP(x)dxdx+C]y = e^{-\int P(x) \, dx} \left[ \int Q(x) e^{\int P(x) dx} dx + C \right]

線性齊次微分方程式 公式推導

由於 Q(x)=0Q(x) = 0

y+P(x)y=0y' + P(x)y = 0

用分離變數法

dydx=P(x)ydyy=P(x)dxln(y)=P(x)dx+Cy=CeP(x)dx\begin{align} \frac{dy}{dx} &= - P(x) y \\ \frac{dy}{y} &= - P(x) dx \\ ln(y) &= - \int P(x) dx + C \\ y &= Ce^{- \int P(x) dx} \end{align}

線性非齊次微分方程式 公式推導

y+P(x)y=Q(x)dydx=Q(x)P(x)y(P(x)yQ(x))dx+dy=0\begin{align} y' + P(x)y &= Q(x) \\ \frac{dy}{dx} &= Q(x) - P(x) y \\ (P(x) y - Q(x))dx + dy &= 0 \end{align} My=P(x)Nx=0MyNx\begin{align} \frac{\partial M}{\partial y} &= P(x) \\ \frac{\partial N}{\partial x} &= 0 \\ \frac{\partial M}{\partial y} &\not = \frac{\partial N}{\partial x} \end{align}

無法使用 Exact Equations 恰當方程式,需要求 Integrating Factors 積分因子 μ(x)\mu(x)

1N(MyNx)=11(P(x)0)=P(x)\begin{align} &\frac{1}{N}(\frac{\partial M}{\partial y} - \frac{\partial N}{\partial x}) \\ & = \frac{1}{1} (P(x) - 0) \\ & = P(x) \end{align} μ(x)=eP(x)dx\mu(x) = e^{\int P(x) dx} eP(x)dx[P(x)yQ(x)]dx+eP(x)dxdy=0e^{\int P(x) dx} \left[ P(x)y - Q(x) \right] dx + e^{\int P(x) dx} dy = 0

轉變為恰當方程式,求 Potential Function 位勢函數:

F=Mdx+k(y)=eP(x)dx[P(x)yQ(x)]dx+k(y)=yeP(x)dxd(P(x)dx)Q(x)eP(x)dxdx+k(y)=yeP(x)dxQ(x)eP(x)dxdx+k(y)F=Ndy+k(x)=eP(x)dxdy+k(x)=yeP(x)dx+k(x)+C\begin{align} F &= \int M dx + k(y) \\ &= \int e^{\int P(x) dx} [ P(x) y - Q(x) ] dx + k(y) \\ &= y \int e^{\int P(x) dx} d (\int P(x) dx) - \int Q(x) e^{\int P(x) dx} dx + k(y) \\ &= y e^{\int P(x) dx} - \int Q(x) e^{\int P(x) dx} dx + k(y) \\ F &= \int N dy + k(x) \\ &= \int e^{\int P(x) dx} d y + k(x) \\ &= y e^{\int P(x) dx} + k(x) + C \end{align}

聯立兩條方程式

{F=yeP(x)dxQ(x)eP(x)dxdx+k(y)F=yeP(x)dx+k(x)+C\left\{ \begin{matrix} F &=& y e^{\int P(x) dx} - \int Q(x) e^{\int P(x) dx} dx + k(y)\\ F &=& y e^{\int P(x) dx} + k(x) + C \end{matrix} \right.

從第二條方程式可知,F 不含 y 變量,故 k(y) = 0

比較兩式, k(y)=Q(x)eP(x)dxdx+Ck(y) = - \int Q(x) e^{\int P(x) dx} dx + C

F=CyeP(x)dxQ(x)eP(x)dxdx+C=CyeP(x)dx=Q(x)eP(x)dxdx+C\begin{align} F &= C \\ y e^{\int P(x) dx} - \int Q(x) e^{\int P(x) dx} dx + C &= C \\ y e^{\int P(x) dx} = \int Q(x) e^{\int P(x) dx} dx + C \end{align} y=eP(x)dx[Q(x)eP(x)dxdx+C]y = e^{- \int P(x) dx} \left[ \int Q(x) e^{\int P(x) dx} dx + C \right]