常數變易法的思想來源是什麼?

這個方法在大部分教材上都沒有特別說明。


方程

a(x)frac{dy}{dx} +b(x)y=c(x)

叫做一階線性微分方程.改寫為

frac{dy}{dx} +frac{b(x)}{a(x)} y=frac{c(x)}{a(x)},記為frac{dy}{dx} +P(x)y=Q(x),或frac{dy}{dx} =-P(x)+Q(x),一般形式為

y=f(x,y).

c(x)=0時,方程a(x)frac{dy}{dx} +b(x)y=0稱為一階齊次線性微分方程,其中a(x)b(x)是自變數

x的函數,這個方程的解可以通過積分求出.

等號兩端同時乘以dx,再除以a(x)y,整理得

frac{1}{y} dy=-frac{b(x)}{ a(x})dx,兩端積分,得

ln |y| =  - int {frac{{b(x)}}{{a(x)}}dx}  + K,其中K是任意常數.則

y =  pm {e^K}{e^{ - int {[b(x)/a(x)]dx} }}

因為K是任意常數,從而pm {e^K}是非零的任意常數,可知y=0也是原方程的解,因此方程的解為

y = C{e^{ - int {[b(x)/a(x)]dx} }},其中C為任意常數.

以上解法只適用於一階齊次線性微分方程,當c(x) 
e 0時,方程

a(x)frac{{dy}}{{dx}} + b(x)y = c(x)

的解會稍微複雜.我們來解這個方程,對於方程的左端,如果b(x)恰好是a(x)的導數,即

a (x)=b(x),那麼左端就是a(x)y的導數,即

frac{{{	ext{d}}[aleft( x 
ight)y]}}{{{	ext{d}}x}} = aleft( x 
ight)frac{{dy}}{{dx}} + aleft( x 
ight)y,

frac{{d[a(x)y]}}{{dx}} = c(x)

兩端積分,得

a(x)y = int {c(x)} dx + K,其中K是任意常數,得

y = frac{1}{{a(x)}}[int {c(x)} dx + K]

上面的解法只適用於y的係數b(x)恰好是frac{{dy}}{{dx}}的係數的導數的情況,這不是一般情況,但是我們也許可以在一般的非齊次線性微分方程兩端乘以某個函數u(x),而將該方程轉化為上面的情況,即

u(x)a(x)frac{{dy}}{{dx}} + u(x)b(x)y = u(x)c(x)

這裡選擇的函數u(x)能夠使u(x)b(x)u(x)a(x)的導數,即

frac{{d[u(x)a(x)]}}{{dx}} = u(x)frac{{da(x)}}{{dx}} + frac{{du(x)}}{{dx}}a(x) = u(x)b(x)

重新整理,得

a(x)frac{{du(x)}}{{dx}} + u(x)[frac{{da(x)}}{{dx}} - b(x)] = 0

這是一個關於函數u(x)的一階齊次線性微分方程,如果我們把

frac{{da(x)}}{{dx}} - b(x)

看作函數u(x)的係數,則

frac{{du(x)}}{{u(x)}} =  - frac{1}{{a(x)}}[frac{{da(x)}}{{dx}} - b(x)]dx

兩端積分,得

int {frac{{du(x)}}{{u(x)}}}  =  - int {frac{1}{{a(x)}}[frac{{da(x)}}{{dx}} - b(x)]dx}

ln |u(x)| =  - int {{ frac{1}{{a(x)}}[frac{{da(x)}}{{dx}} - b(x)]} dx}  + K

其中K是任意常數.

left| {u(x)} 
ight| = {{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}{	ext{d}}x}  + K}}

u(x) =  pm {{	ext{e}}^K} cdot {{	ext{e}}^{ - mathop {int {left{ {frac{1}{{a(x)}}left[ {frac{{{	ext{d}}a(x)}}{{{	ext{d}}x}} - b(x)} 
ight]} 
ight}{	ext{d}}x} }
olimits }}

其中pm ±e^K

是不為零的任意常數,因為u(x)=0也使方程

a(x)frac{{du(x)}}{{dx}} + u(x)[frac{{da(x)}}{{dx}} - b(x)] = 0

成立,因此用任意常數k代替pm ±e^K,解得

u(x) = k{{	ext{e}}^{ - int {left{ {frac{1}{{a(x)}}left[ {frac{{{	ext{d}}a(x)}}{{{	ext{d}}x}} - b(x)} 
ight]} 
ight}{	ext{d}}x} }}

從而,一階非齊次線性微分方程a(x)frac{{dy}}{{dx}} + b(x)y = c(x)在乘以函數u(x)後所得方程

u(x)a(x)frac{{dy}}{{dx}} + u(x)b(x)y = u(x)c(x)

的解可以以如下形式表示

d{ [u(x)a(x)]y}  = u(x)c(x)dx

兩端積分,得

u(x)a(x)y = int {u(x)c(x)dx}  + C

y = frac{1}{{u(x)a(x)}}[int {u(x)c(x)dx}  + C

y = frac{1}{{kaleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}}}left[ {int {kcleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}} dx + C} 
ight]

y = frac{1}{{aleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}}}left[ {int {cleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}} dx + frac{C}{k}} 
ight]

y = frac{1}{{aleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}}}left[ {int {cleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}} dx + C} 
ight]

y = frac{1}{{aleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}}}left[ {int {cleft( x 
ight){{	ext{e}}^{ - int {left{ {frac{1}{{aleft( x 
ight)}}left[ {frac{{{	ext{d}}aleft( x 
ight)}}{{{	ext{d}}x}} - bleft( x 
ight)} 
ight]} 
ight}} {	ext{d}}x}}} dx + C} 
ight]

對於非齊次一階線性微分方程frac{{dy}}{{dx}} + P(x)y = Q(x)

兩端乘以u(x),得

u(x)frac{{dy}}{{dx}} + u(x)P(x)y = u(x)Q(x)

frac{{du(x)}}{{dx}} = u(x)P(x)

frac{{du(x)}}{{u(x)}} = P(x){	ext{d}}x

ln |u(x)| = int {P(x){	ext{d}}x}  + C

u(x) =  pm {{	ext{e}}^C}{{	ext{e}}^{int {P(x){	ext{d}}x} }}

u(x) = {C_1}{{	ext{e}}^{int {P(x){	ext{d}}x} }}

frac{{d[u(x)y]}}{{{	ext{d}}x}} = u(x)Q(x)

d[u(x)y] = u(x)Q(x){	ext{d}}x

u(x)y = int {u(x)Q(x){	ext{d}}x}  + {C_2}

y = frac{1}{{u(x)}}[int {u(x)Q(x){	ext{d}}x}  + {C_2}]

y = frac{1}{{{C_1}{{	ext{e}}^{int {P(x){	ext{d}}x} }}}}[int {{C_1}{{	ext{e}}^{int {P(x){	ext{d}}x} }}Q(x){	ext{d}}x}  + {C_2}]

y = frac{1}{{{C_1}{{	ext{e}}^{int {P(x){	ext{d}}x} }}}}int {{C_1}{{	ext{e}}^{int {P(x){	ext{d}}x} }}Q(x){	ext{d}}x}  + frac{{{C_2}}}{{{C_1}{{	ext{e}}^{int {P(x){	ext{d}}x} }}}}

y = frac{1}{{{{	ext{e}}^{int {P(x){	ext{d}}x} }}}}int {{{	ext{e}}^{int {P(x){	ext{d}}x} }}Q(x){	ext{d}}x}  + C{{	ext{e}}^{ - int {P(x){	ext{d}}x} }}

y = {{	ext{e}}^{ - int {P(x){	ext{d}}x} }}int {{{	ext{e}}^{int {P(x){	ext{d}}x} }}Q(x){	ext{d}}x}  + C{{	ext{e}}^{ - int {P(x){	ext{d}}x} }}

y = C{{	ext{e}}^{ - int {P(x){	ext{d}}x} }} + {{	ext{e}}^{ - int {P(x){	ext{d}}x} }}int {Q(x){{	ext{e}}^{int {P(x){	ext{d}}x} }}{	ext{d}}x}

來源:阿爾法微積分,專註微積分

轉載請註明出處。


學這塊時我的理解是根據f(x)×g(x)的求導法則想到常數變異的。

[如果這個問題問的是前人如何想到常數變異的話]


淺析常微分方程的常數變易法_百度文庫


推薦閱讀:

數學物理方法和微分方程學科的區別是什麼?
微分兩邊能不能同時做定積分?
三階微分方程的初值問題的幾何意義是什麼?
現行微積分體系的錯誤:dx與Δx的關係幽靈?
一階線性齊次微分方程中的齊次的含義?

TAG:微積分 | 高等數學 | 微分方程 |