Chapter 9 Solutions
1.以傅立葉轉換(Fourier Transform)解下列常微分方程式
𝑦 ′′ − 𝑘 2 𝑦 = 𝑒 −𝑥 , 𝑦 ′ (0) = 0
[Sol]
等號兩邊同取 Fourier cosine transform
𝐹𝑐 [𝑦 ′′ − 𝑘 2 𝑦] = 𝐹𝑐 (𝑒 −𝑥 )
1
[−𝜔2 𝐹𝑐 (𝑦) − 𝑦’(0)] − 𝑘 2 𝐹𝑐 (𝑦) = 公式: 𝐹𝑐 ( 𝑓”) = −𝜔2 𝐹𝑐 (𝑓) − 𝑓’(0)
𝜔2 +1
1
(−𝜔2 − 𝑘 2 )𝐹𝑐 (𝑦) =
𝜔2 +1
−1
𝐹𝑐 (𝑦) =
(𝜔 2 + 1)(𝜔 2 + 𝑘 2 )
2 ∞
⇒ 𝑦 (𝑥) = ∫ 𝐹 (𝑦)𝑐𝑜𝑠𝜔𝑥 𝑑𝜔
𝜋 0 𝑐
−2 ∞ 1
= ∫ 𝑐𝑜𝑠𝜔𝑥𝑑𝜔 ##
𝜋 0 (𝜔 + 1)(𝜔 2 + 𝑘 2 )
2
1
2.以 Fourier Transform 解𝑢(𝑥, 𝑡)。
𝜕𝑢 𝜕 2𝑢
=𝛼 2
𝜕𝑡 𝜕𝑥
𝜕𝑢
| = 𝑢0
𝜕𝑥 𝑥 = 0
𝑢(∞, 𝑡) = 𝑏𝑜𝑢𝑛𝑑𝑒𝑑
𝑢(𝑥, 0) = 0
[Sol]
Take Fourier Cosine Transform with respect to x
𝜕𝑢 𝜕 2𝑢
𝐹𝑐 ( ) = 𝐹𝑐 (𝛼 2 )
𝜕𝑡 𝜕𝑥
𝜕
⇒ 𝐹 (𝑢) = 𝛼[−𝜔2 𝐹𝑐 (𝑢) − 𝑢′ (0, 𝑡)]
𝜕𝑡 𝑐
= −𝛼𝜔2 𝐹𝑐 (𝑢) − 𝛼𝑢0
∧
𝑆𝑒𝑡 𝐹𝑐 (𝑢) = 𝑢𝑐 (𝜔, 𝑡)
∧
𝜕𝑢𝑐 ∧
⇒ + 𝛼𝜔2 𝑢𝑐 = −𝛼𝑢0
𝜕𝑡
2 𝑑𝑡 2𝑡
𝐼𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑛𝑔 𝑓𝑎𝑐𝑡𝑜𝑟(𝐼𝐹) = 𝑒 ∫ 𝛼𝜔 = 𝑒 𝛼𝜔
∧
𝜕𝑢𝑐 ∧ 2 2
⇒[ + 𝛼𝜔2 𝑢𝑐 ]𝑒 𝛼𝜔 𝑡 = (−𝛼𝑢0 ) 𝑒 𝛼𝜔 𝑡
𝜕𝑡
∧ 2
𝑑(𝑢𝑐 𝑒 𝛼𝜔 𝑡 ) 2
= (−𝛼𝑢0 )𝑒 𝛼𝜔 𝑡
𝑑𝑡
∧ 2𝑡 2 −𝛼𝑢0 𝛼𝜔2𝑡
𝑢𝑐 𝑒 𝛼𝜔 = (−𝛼𝑢0 ) ∫ 𝑒 𝛼𝜔 𝑡 𝑑𝑡 = 𝑒 + 𝑐(𝜔)
𝛼𝜔 2
∧ −𝑢0 2
𝑢𝑐 (𝜔, 𝑡) = 2
+ 𝑐(𝜔)𝑒 −𝛼𝜔 𝑡
𝜔
2
𝑢(𝑥, 0) = 0
∧ −𝑢0 𝑢0
𝑢𝑐 (𝜔, 0) = 0 = 2
+ 𝑐(𝜔) ⇒ 𝑐(𝜔) = 2
𝜔 𝜔
∧ −𝑢0 −𝛼𝜔 2 𝑡
−𝑢0 𝑢0 −𝛼𝜔2𝑡
𝑢𝑐 (𝜔, 𝑡) = + 𝑐(𝜔)𝑒 = + 𝑒
𝜔2 𝜔2 𝜔2
2 ∞∧ 2 ∞ −𝑢0 𝑢0 2
𝑢(𝑥, 𝑡) = ∫ 𝑢𝑐 (𝜔, 𝑡) 𝑐𝑜𝑠𝜔𝑥𝑑𝜔 = ∫ ( 2 + 2 𝑒 −𝛼𝜔 𝑡 ) 𝑐𝑜𝑠𝜔𝑥𝑑𝜔 ##
𝜋 0 𝜋 0 𝜔 𝜔
3
𝜕𝜙 𝜕 2 𝜙
3. = , 𝜙(𝑥, 𝑡) =?
𝜕𝑡 𝜕𝑥 2
3 ,0 ≤ 𝑡 ≤ 2
𝜙(0, 𝑡) = {
0, 𝑡>2
𝜙(∞, 𝑡) = 𝑏𝑜𝑢𝑛𝑑𝑒𝑑
𝜙(𝑥, 0) = 2
1 𝑎
ℒ −1 ( 𝑒 −𝑎√𝑠 ) = 𝑒𝑟𝑓𝑐( )
𝑠 2√𝑡
[Sol]
對"t"取 Laplace Transform !
𝜕𝜙 𝜕2𝜙
ℒ( ) = ℒ( 2 )
𝜕𝑡 𝜕𝑥
𝜕2
⇒ 𝑠ℒ[𝜙] − 𝜙(𝑥, 0) = ℒ[𝜙]
𝜕𝑥 2
𝜕2
𝑠ℒ[𝜙] − 2 = ℒ[𝜙] ( ∵ 𝝓(𝒙, 𝟎) = 𝟐 )
𝜕𝑥 2
𝜕2
⇒ ℒ[𝜙] − 𝑠ℒ[𝜙] = −2
𝜕𝑥 2
𝐿𝑒𝑡 ℒ[𝜙] = 𝑄(𝑥, 𝑠)
𝜕2
∴ 2 ℒ[𝜙] − 𝑠ℒ[𝜙] = −2
𝜕𝑥
𝜕2
⇒ 2 𝑄 − 𝑠𝑄 = −2
𝜕𝑥
⇒ 𝑄" − 𝑠𝑄 = −2 (2nd order of Q(x, s)𝑤𝑖𝑡ℎ 𝑟𝑒𝑠𝑝𝑒𝑐𝑡 𝑡𝑜 𝑥)
𝑄(𝑥, 𝑠) =? ?
Q" − 𝑠𝑄 = 0 ⇒ 𝑄ℎ
Solve {
𝑄" − 𝑠𝑄 = −2 ⇒ 𝑄𝑝
∵ 𝑄ℎ " − 𝑠𝑄ℎ = 0
4
Let 𝑄ℎ = 𝑒 𝑎𝑥
∴ 𝑄ℎ " = 𝑎2 𝑒 𝑎𝑥
𝑄ℎ " − 𝑠𝑄𝐻 = 0
⇒ 𝑎2 𝑒 𝑎𝑥 − 𝑠𝑒 𝑎𝑥 = 0
⇒ (𝑎2 − 𝑠)𝑒 𝑎𝑥 = 0
∴ 𝑎 = ±√𝑠
∴ 𝑄ℎ = 𝑐1 (𝑠)𝑒 √𝑠𝑥 + 𝑐2 (𝑠)𝑒 −√𝑠𝑥
∵ 𝑄𝑝 " − 𝑠𝑄𝑝 = −2
Let 𝑄𝑃 = 𝐴
∴ 𝑄𝑝 " = 0
∴ 𝑄𝑝 " − 𝑠𝑄𝑃 = −2
⇒ 0 − 𝑠𝐴 = −2
2
∴𝐴= = 𝑄𝑝
𝑠
2
∴ 𝑄 = 𝑄ℎ + 𝑄𝑝 = 𝑐1 (𝑠)𝑒 √𝑠𝑥 + 𝑐2 (𝑠)𝑒 −√𝑠𝑥 +
𝑠
∵ 𝜙(∞, 𝑡) = 𝑏𝑜𝑢𝑛𝑑𝑒𝑑
2 2
∴ ℒ[𝜙(∞, 𝑡)] = 𝑄(∞, 𝑠) = 𝑏𝑜𝑢𝑛𝑑𝑒𝑑 = 𝑐1 (𝑠)𝑒 √𝑠∞ + 𝑐2 (𝑠)𝑒 −√𝑠∞ + = 𝑐1 (𝑠)𝑒 √𝑠∞ +
𝑠 𝑠
∴ 𝑐1 (𝑠) = 0
3 ,0 ≤ 𝑡 ≤ 2
∵ 𝜙(0, 𝑡) = {
0,𝑡 > 2
∞
∴ 𝑄(0, 𝑠) = ℒ[𝜙(0, 𝑡)] = ∫ 𝜙(0, 𝑡)𝑒 −𝑠𝑡 𝑑𝑡
0
5
2 ∞ 2
= ∫ ( 3)𝑒 −𝑠𝑡 𝑑𝑡 + ∫ (0)𝑒 −𝑠𝑡 𝑑𝑡 = ∫ ( 3)𝑒 −𝑠𝑡 𝑑𝑡
0 2 0
−3 −𝑠𝑡 2
= 𝑒 │
𝑠 0
−3 −2𝑠 2 2
= (𝑒 − 1) = 𝑐2 (𝑠)𝑒 −√𝑠∙0 + = 𝑐2 (𝑠) +
𝑠 𝑠 𝑠
−3 −2𝑠 1
∴ 𝑐2 (𝑠) = 𝑒 +
𝑠 𝑠
2 −3 1 2
∴ 𝑄(𝑥, 𝑠) = 𝑐1 (𝑠)𝑒 √𝑠𝑥 + 𝑐2 (𝑠)𝑒 −√𝑠𝑥 + = ( 𝑒 −2𝑠 + )𝑒 −√𝑠𝑥 +
𝑠 𝑠 𝑠 𝑠
−3 −2𝑠 −√𝑠𝑥 1 −√𝑠𝑥 2
= 𝑒 𝑒 + 𝑒 +
𝑠 𝑠 𝑠
−3 −2𝑠 −√𝑠𝑥 1 2
𝜙(𝑥, 𝑡) = ℒ −1 [𝑄(𝑥, 𝑠)] = ℒ −1 ( 𝑒 𝑒 ) + ℒ −1 ( 𝑒 −√𝑠𝑥 ) + ℒ −1 ( )
𝑠 𝑠 𝑠
−3 −𝑥√𝑠 −2𝑠 1 2
= ℒ −1 [( 𝑒 )𝑒 ] + ℒ −1 ( 𝑒 −𝑥√𝑠 ) + ℒ −1 ( )
𝑠 𝑠 𝑠
𝑥 𝑥
= −3 𝑒𝑟𝑓𝑐( )𝑢(𝑡 − 2) + 𝑒𝑟𝑓𝑐( ) + 2 ##
2√𝑡 − 2 2√𝑡
𝑢(𝑡 − 2): 𝑢𝑛𝑖𝑡 𝑠𝑡𝑒𝑝 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
(ℒ −1 [F(𝑠)𝑒 −𝑎𝑠 ] = 𝑓(𝑡 − 𝑎) 𝑢(𝑡 − 𝑎))
6
4.以 Laplace Transform 解𝑤(𝑥, 𝑡)。
𝜕 2𝑤 𝜕 2𝑤
= 2 2 , 𝑥 ≥ 0, 𝑡 ≥ 0
𝜕𝑡 2 𝜕𝑥
𝑤(0, 𝑡) = 𝑐𝑜𝑠𝜋𝑡
𝑤(∞, 𝑡) = 𝑏𝑜𝑢𝑛𝑑𝑒𝑑
𝑤(𝑥, 0) = 1 + 𝑒 −𝑥
𝜕𝑤
| =0
𝜕𝑡 𝑡 = 0
[Sol]
Take Laplace Transform with respect to t
𝜕 2𝑤 𝜕 2𝑤
ℒ( ) = 2ℒ( )
𝜕𝑡 2 𝜕𝑥 2
𝜕2
⇒ 𝑠 2 ℒ(𝑤) − 𝑠𝑤(𝑥, 0) − 𝑤 ′ (𝑥, 0) = 2 ℒ(𝑤)
𝜕𝑥 2
𝜕 2𝑊 𝑠2 −𝑠 𝑠 −𝑥
− 𝑊 = − 𝑒
𝜕𝑥 2 2 2 2
𝜕 2𝑊 𝑠2
− 𝑊 = 0 ⇒ 𝑊ℎ (𝑥, 𝑠)
𝜕𝑥 2 2
𝜕 2𝑊 𝑠2 −𝑠 𝑠 −𝑥
− 𝑊 = − 𝑒 ⇒ 𝑊𝑝 (𝑥, 𝑠)
𝜕𝑥 2 2 2 2
𝑠 −𝑠
𝑥 𝑥
𝑊ℎ (𝑥, 𝑠) = 𝑐1 (𝑠)𝑒 √2 + 𝑐2 (𝑠)𝑒 √2
𝑆𝑒𝑡 𝑊𝑝 (𝑥, 𝑠) = 𝐴𝑒 −𝑥 + 𝐵, 𝑊𝑝 ′ = −𝐴𝑒 −𝑥 , 𝑊𝑝 " = 𝐴𝑒 −𝑥
𝜕 2𝑊 𝑠2 −𝑠 𝑠 −𝑥
代回 2
− 𝑊= − 𝑒
𝜕𝑥 2 2 2
7
−𝑥
𝑠2 −𝑠 𝑠 −𝑥
𝐴𝑒 − (𝐴𝑒 −𝑥 + 𝐵) = − 𝑒
2 2 2
𝐵𝑠 2 𝑠2 −𝑠 𝑠 −𝑥
− + 𝐴(1 − )𝑒 −𝑥 = − 𝑒
2 2 2 2
𝐵𝑠 2 −𝑠
− =
2 2
𝑠2 −𝑠
𝐴(1 − )=
2 2
𝑠2 −𝑠
𝐴(1 − ) =
2 2
𝑠 1
𝐴= ,𝐵 =
𝑠2 −2 𝑠
𝑠 1
𝑊𝑝 (𝑥, 𝑠) = 𝐴𝑒 −𝑥 + 𝐵 = 𝑒 −𝑥 +
𝑠2 −2 𝑠
𝑠 −𝑠
𝑥 𝑥 𝑠 1
𝑊(𝑥, 𝑠) = 𝑊ℎ (𝑥, 𝑠) + 𝑊𝑝 (𝑥, 𝑠) = 𝑐1 (𝑠)𝑒 √2 + 𝑐2 (𝑠)𝑒 √2 + 𝑒 −𝑥 +
𝑠2 −2 𝑠
𝑤(∞, 𝑡) = 𝑏𝑜𝑢𝑛𝑑𝑒𝑑
𝑠 −𝑠
∞ ∞ 𝑠 −∞
1
𝑊(∞, 𝑠) = 𝑏𝑜𝑢𝑛𝑑𝑒𝑑 = 𝑐1 (𝑠)𝑒 √2 + 𝑐2 (𝑠)𝑒 √2 + 𝑒 +
𝑠2 − 2 𝑠
𝑐1 (𝑠) = 0
𝑤(0, 𝑡) = 𝑐𝑜𝑠𝜋𝑡
𝑠 𝑠 1
𝑊(0, 𝑠) = = 𝑐2 (𝑠) + +
𝑠2 + 𝜋2 𝑠2 − 2 𝑠
𝑠 𝑠 1
𝑐2 (𝑠) = − 2 −
𝑠2 +𝜋 2 𝑠 −2 𝑠
8
𝑠 𝑠 1 −𝑠𝑥 𝑠 1
𝑊(𝑥, 𝑠) = ( 2 2
− 2
− )𝑒 √2 + 2 𝑒 −𝑥 +
𝑠 +𝜋 𝑠 −2 𝑠 𝑠 −2 𝑠
𝑠 𝑠 1 −𝑥𝑠 𝑠 1
=( 2 2
− 2
− )𝑒 √2 + 2 𝑒 −𝑥 +
𝑠 +𝜋 𝑠 −2 𝑠 𝑠 −2 𝑠
1 1 1 1
𝑠 2 2 1 −𝑥𝑠 2 2 )𝑒 −𝑥 + 1
=( 2 − ( + ) − )𝑒 √2 + ( +
𝑠 + 𝜋2 𝑠 + √2 𝑠 − √2 𝑠 𝑠 + √2 𝑠 − √2 𝑠
𝑤(𝑥, 𝑡) = ℒ −1 𝑊(𝑥, 𝑠)
1 1 −𝑥
1 1
𝑠 2 2 1 𝑠 2 2 )𝑒 −𝑥 + 1]
= ℒ −1 [( 2 − ( + ) − )𝑒 √2 + ( +
𝑠 + 𝜋2 𝑠 + √2 𝑠 − √2 𝑠 𝑠 + √2 𝑠 − √2 𝑠
𝑥 1 −√2(𝑡− 𝑥 ) 𝑥
√2(𝑡− ) 𝑥 1
= [𝑐𝑜𝑠𝜋 (𝑡 − ) − (𝑒 √2 + 𝑒 √2 ) − 1]𝑢(𝑡 − ) + 𝑒 −𝑥 [ (𝑒 −√2𝑡 + 𝑒 √2𝑡 )] + 1
√2 2 √2 2
𝑥
𝑐𝑜𝑠ℎ(√2 (𝑡 − )) 𝑐𝑜𝑠ℎ(√2𝑡)
√2
𝑥
𝑢 (𝑡 − ) : 𝑢𝑛𝑖𝑡 𝑠𝑡𝑒𝑝 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 ##
√2