Assignment 2 Due date: July 21, 2015 For Problem 1, 2, 3, 4, 5 and 6, you only need to three of them. 1. For the differential equation 2x2 y 00 + (3x − x2 )y 0 − y = 0 for x > 0, find the first three terms of a non-zero series solution at x = 0, satisfying lim y(x) = 0. x&0 Answer. Since 2x2 y 00 + (3x − x2 )y 0 − y = 0, then y 00 + 3x − x2 0 1 y − 2 y = 0. 2 2x 2x That is, y 00 + Let p(x) = 3−x 0 1 y − 2 y = 0. 2x 2x 1 3−x and q(x) = − 2 , then 2x 2x xp(x) = 3−x 3 1 = − x, 2 2 2 1 and x2 q(x) = − . 2 So x = 0 is a regular singular point to the differential equation 2x2 y 00 + (3x − x2 )y 0 − y = 0, and the indicial equation is: 3 1 1 1 0 = r(r − 1) + p0 r + q0 = r(r − 1) + r − = r2 + r − . 2 2 2 2 So the exponents at x = 0 are: 1 r1 = , 2 Let φ(x, r) = xr ∞ X k=0 ak (r)xk = ∞ X and r2 = −1. ak (r)xk+r for some {ak (r)}∞ k=0 which will be determined later, then k=0 φx (x, r) = ∞ X (k + r)ak (r)xk+r−1 k=0 φxx (x, r) = ∞ X k=0 (k + r)(k + r − 1)ak (r)xk+r−2 . So we have 2x2 φxx (x, r) + (3x − x2 )φx (x, r) − φ(x, r) ∞ ∞ ∞ X X X = 2x2 · (k + r)(k + r − 1)ak (r)xk+r−2 + (3x − x2 ) (k + r)ak (r)xk+r−1 − ak (r)xk+r k=0 = = ∞ X k=0 ∞ X 2(k + r)(k + r − 1)ak (r)xk+r + (k + r)ak (r)xk+r+1 − k=0 k=0 k=0 ∞ X ∞ X ∞ X 2(k + r)(k + r − 1)ak (r)xk+r + k=0 = 3(k + r)ak (r)xk+r − k=0 ∞ X ∞ X 2(n + r)(n + r − 1)an (r)xn+r + n=0 3(k + r)ak (r)xk+r − k=0 ∞ X n=0 = 2r(r − 1)a0 (r) + 3rao (r) − a0 (r) + ∞ X ak (r)xk+r k=0 (n + r − 1)an−1 (r)xn+r − n=1 ∞ X 3(n + r)an (r)xn+r − ∞ X ∞ X ak (r)xk+r k=0 ∞ X (n + r − 1)an−1 (r)xn+r − n=1 an (r)xn+r n=0 [2(n + r)(n + r − 1)an (r) + 3(n + r)an (r) − (n + r − 1)an−1 (r) n=1 −an (r)]xn+r 2 = (2r − 2r + 3r − 1)a0 (r) + ∞ X {[2(n + r)(n + r − 1) + 3(n + r) − 1]an (r) − (n + r − 1)an−1 (r)}xn+r n=1 = (2r2 + r − 1)a0 (r) + ∞ X {[2(n + r)(n + r − 1) + 3(n + r) − 1]an (r) − (n + r − 1)an−1 (r)}xn+r . n=1 Let a0 (r) = 1, and define the sequence {an (r)}∞ n=1 be the following recurrence relation: [2(n + r)(n + r − 1) + 3(n + r) − 1]an (r) − (n + r − 1)an−1 (r) = 0, That is, an (r) = = n+r−1 an−1 (r) 2(n + r)(n + r − 1) + 3(n + r) − 1 n+r−1 an−1 (r), ∀n ≥ 1. (n + r)(2n + 2r + 1) − 1 So we get 2x2 φxx (x, r) + (3x − x2 )φx (x, r) − φ(x, r) = 2r2 + r − 1. Page 2 ∀n ≥ 1. ∞ 1 X 1 1 2 =x xk , then lim y(x) = 0 and Let y(x) = φ x, ak x&0 2 2 k=0 1 a0 = 1 2 1 + 12 − 1 1 1 = a1 · a0 1 1 2 2 (1 + 2 )(2 · 1 + 2 · 2 + 1) − 1 1 = 10 2 + 12 − 1 1 1 a2 = · a1 1 1 2 2 (2 + 2 )(2 · 2 + 2 · 2 + 1) − 1 3 = . 280 2. Consider the differential equation 2(x − 1)y 00 + y 0 + y = 0 for x > 1. (a) Show that x = 1 is a regular singular point. Find the indicial equation and the exponents at x = 1. Answer. Since 2(x − 1)y 00 + y 0 + y = 0, then y 00 + Let p(x) = 1 1 y0 + y = 0. 2(x − 1) 2(x − 1) 1 1 and q(x) = , then 2(x − 1) 2(x − 1) 1 (x − 1)p(x) = , 2 So x = 1 is a regular singular point and p0 = and 1 2 (x − 1)2 q(x) = x−1 . 2 and q0 = 0. Then the indicial equation is: 1 1 0 = r(r − 1) + p0 r + q0 = r(r − 1) + r = r2 − r. 2 2 Then the exponents at x = 1 are r1 = 1 2 and r2 = 0. (b) Find the first two non-zero terms of two linearly independent series solutions about x = 1. Answer. r Let φ(x, r) = (x − 1) ∞ X k ak (r)(x − 1) = k=0 ∞ X ak (r)(x − 1)k+r for some {ak (r)}∞ k=0 which will k=0 be determined later, then φx (x, r) = ∞ X (k + r)ak (r)(x − 1)k+r−1 k=0 φxx (x, r) = ∞ X (k + r)(k + r − 1)ak (r)(x − 1)k+r−2 . k=0 Page 3 Then we have 2(x − 1)φxx (x, r) + φx (x, r) + φ(x, r) ∞ ∞ ∞ X X X = 2(x − 1) (k + r)(k + r − 1)ak (r)(x − 1)k+r−2 + (k + r)ak (r)(x − 1)k+r−1 + ak (r)(x − 1)k+r k=0 = = ∞ X k=0 2(k + r)(k + r − 1)ak (r)(x − 1)k+r−1 + ∞ X k=0 (k + r)ak (r)(x − 1)k+r−1 + ∞ X k=0 k=0 k=0 ∞ X ∞ X ∞ X 2(k + r)(k + r − 1)ak (r)(x − 1)k+r−1 + k=0 (k + r)ak (r)(x − 1)k+r−1 + k=0 r−1 r−1 ak (r)(x − 1)k+r ak−1 (r)(x − 1)k+r−1 k=1 = 2r(r − 1)a0 (r)(x − 1) + ra0 (r)(x − 1) ∞ X + [2(k + r)(k + r − 1)ak (r) + (k + r)ak (r) + ak−1 (r)](x − 1)k+r−1 k=1 = [2r(r − 1) + r]a0 (r)(x − 1)r−1 + ∞ X {(k + r)[2(k + r − 1) + 1]ak (r) + ak−1 (r)}(x − 1)k+r−1 k=1 = [2r2 − r]a0 (r)(x − 1)r−1 + ∞ X [(k + r)(2k + 2r − 1)ak (r) + ak−1 (r)](x − 1)k+r−1 . k=1 Let a0 (r) = 0, define {ak (r)}∞ k=1 be the following recurrence relation: (k + r)(2k + 2r − 1)ak (r) + ak−1 (r) = 0, for all k ≥ 1. Then we get ak (r) = − ak−1 (r) , (k + r)(2k + 2r − 1) for all k ≥ 1. So we have 2(x − 1)φxx (x, r) + φx (x, r) + φ(x, r) = (2r2 − r)(x − 1)r−1 . Let X ∞ 1 1 1 y1 (x) = φ x, = ak (x − 1)k+ 2 , 2 2 k=0 and y2 (x) = φ(x, 0) = ∞ X ak (0)(x − 1)k . k=0 Since a0 (r) = 1, then y1 (x) and y2 (x) are two linearly independent solutions. Notice that 1 a0 = a0 (0) = 1 2 a0 12 1 a1 = − 2 1 + 12 2 · 1 + 2 · 12 − 1 1 = − 3 a0 (0) a1 (0) = − (1 + 0)(2 · 1 + 2 · 0 − 1) = −1. Page 4 (c) Find the first two non-zero terms of the series solution about x = 1, satisfying lim y(x) = 1 and x&1 √ lim x − 1y 0 (x) = 1. x&1 Answer. By the result of part (b), then there are two constants C1 and C2 such that y(x) = C1 y1 (x) + C2 y2 (x) 1 3 1 2 2 = C1 (x − 1) − (x − 1) + · · · + C2 1 − (x − 1) + a2 (0)(x − 1)2 + · · · . 3 Since lim y(x) = 1, then C2 = 1. Since x&1 0 y (x) = C1 1 1 3 1 − 12 2 (x − 1) − · (x − 1) + · · · + C2 [−1 + 2a2 (x − 1) + · · · ] . 2 2 3 Then √ 0 x − 1y (x) = C1 √ i h √ 1 1 1 3 1 +1 − · (x − 1) 2 + · · · + C2 − x − 1 + 2a2 (x − 1)1+ 2 + · · · . 2 2 3 C1 = 1, that is, C1 = 2. So we get 2 1 3 1 2 2 y(x) = 2y1 (x) + y2 (x) = 2 (x − 1) − (x − 1) + · · · + 1 − (x − 1) + a2 (0)(x − 1)2 + · · · . 3 Since lim x&1 x − 1y 0 (x) = 1, then 3. Consider the Laguerre differential equation: xy 00 + (1 − x)y 0 + λy = 0, x > 0, where λ is a real constant. (a) Show that x = 0 is a regular singular point. Find the indicial equation and the exponents at x = 0. Answer. Since xy 00 + (1 − x)y 0 + λy = 0, then y 00 + Let p(x) = 1 − x 0 λx y + = 0. x y 1−x λ and q(x) = , then x x xp(x) = 1 − x, and x2 q(x) = λx. So x = 0 is a regular singular point, p0 = 1 and q0 = 0. So the indicial equation is: 0 = r(r − 1) + p0 r + q0 = r(r − 1) + r = r2 − r + r = r2 . Then the exponents at x = 0 are r1 = r2 = 0. Page 5 (b) Find the general series solutions near x = 0. Answer. Let φ(x, r) = x r ∞ X k ak (r)x = k=0 ∞ X ak (r)xk+r for some {ak (r)}∞ k=0 which will be determined k=0 later, then φx (x, r) = ∞ X (k + r)ak (r)xk+r−1 k=0 φxx (x, r) = ∞ X (k + r)(k + r − 1)ak (r)xk+r−2 . k=0 Then xφxx (x, r) + (1 − x)φx (x, r) + λφ(x, r) ∞ ∞ ∞ X X X = x (k + r)(k + r − 1)ak (r)xk+r−2 + (1 − x) (k + r)ak (r)xk+r−1 + λ ak (r)xk+r = = k=0 ∞ X ∞ X k=0 ∞ X k=0 ∞ X k=0 ∞ X k=0 ∞ X k=0 k=1 (k + r)(k + r − 1)ak (r)xk+r−1 + (k + r)(k + r − 1)ak (r)xk+r−1 + k=0 +λ ∞ X k=0 (k + r)ak (r)xk+r−1 − (k + r)ak (r)xk+r−1 − (k + r)ak (r)xk+r + λ (k − 1 + r)ak−1 (r)xk+r−1 ak−1 (r)xk+r−1 k=1 k=1 r−1 = r a0 (r)x + ∞ X [(k + r)(k + r − 1 + 1)ak (r) − (k − 1 + r − λ)ak−1 (r)]xk+r−1 k=1 ∞ X = r2 a0 (r) + [(k + r)2 ak (r) − (k − 1 + r − λ)ak−1 (r)]xk+r−1 . k=1 Let a0 (r) = 1, define {ak (r)}∞ k=1 be the following recurrence relation: (k + r)2 ak (r) − (k − 1 + r − λ)ak−1 (r) = 0, ∀k ≥ 1. That is, ak (r) = k−1+r−λ · ak−1 (r), (k + r)2 Page 6 ak (r)xk+r k=0 = r(r − 1)a0 (r)xr−1 + ra0 (r)xr−1 ∞ X + [(k + r)(k + r − 1)ak (r) + (k + r)ak (r) − (k − 1 + r)ak−1 (r) + λak−1 (r)]xk+r−1 2 ∞ X ∀k ≥ 1. That is, we have ak (r) = = = k−1+r−λ (k + r)2 k−1+r−λ (k + r)2 k−1+r−λ (k + r)2 = a0 (r) k−1−1+r−λ · ak−1−1 (r) (k − 1 + r)2 k−1−1+r−λ 1−1+r−λ · · ··· · a0 (r) (k − 1 + r)2 (1 + r)2 · k Y i−1+r−λ i=1 = · ak−1 (r) (i + r)2 k Y i−1−λ+r i=1 (i + r)2 , for all k ≥ 1. Then xφxx (x, r) + (1 − x)φx (x, r) + λφ(x, r) = r2 . Let y1 (x) = φ(x, 0) = 1 + ∞ X k=1 ∞ k X Y i−1−λ ak (0)xk = i2 k=0 (1) ! xk , then y1 (x) is a solution. Notice that i=1 a0 (0) = 1 1−1−λ+0 a1 (0) = (1 + 0)2 = −λ 2−1+0−λ a2 (0) = · (−λ) (2 + 0)2 1−λ (−λ) = 4 λ2 − λ . = 4 By (1), then we get xφxxr (x, r) + (1 − x)φxr (x, r) + λφr (x, r) = 2r. Let y2 (x) = φr (x, 0) = ∞ X a0k (0)xk + k=0 ∞ X ak (0)xk ln x, then y2 (x) is also a solution. Since a0 (r) = 1, k=0 1−1−λ+r r−λ 2−1+r−λ 1+r−λ r−λ a1 (r) = = , and a2 (r) = · a1 (r) = · , then 2 2 2 2 (1 + r) (r + 1) (2 + r) (r + 2) (r + 1)2 a00 (0) = 0, a01 (0) = 1 + 2λ, Page 7 and a02 (0) = 1 + λ − 3λ2 . 4 Then general solution is : y(x) = C1 y1 (x) + C2 y2 (x) λ2 − λ 2 x + ··· = C1 1 − λx + 4 1 + λ − 3λ2 2 λ2 − λ 2 +C2 (1 + 2λ)x + x + · · · + ln x 1 − λx + x + ··· . 4 4 (c) Show that if λ is a non-negative integer n, then there is a polynomial solution of degree n. Answer. By the result of part (b), then ∞ k X Y i−1−λ y1 (x) = 1 + i2 k=1 ! xk , i=1 is a solution. When λ = 0, it’s easy to see that y = 1 is a solution. Now when λ = n for some positive integer, then k Y i−1−λ i=1 i2 k Y i−1−n = i2 i=1 = 0, for all k ≥ n + 1. On the other hand, we know that an (0) = n Y i−1−λ i=1 i2 = n Y i−1−n i=1 i2 = (−1)n n Y n+1−i i=1 i2 = (−1)n · n! (−1)n = . (n!)2 n! So y1 (x) is a polynomial solution of degree n. (d) Find a polynomial solution for each of the case λ = n = 0, 1, 2, 3. Answer. By the result of part (c), then • When λ = 0, y = 1 is a polynomial solution. • When λ = 1, y = 1 − x is a polynomial solution. 1−1−2 1 1 • When λ = 2, y = 1 + x + x2 = 1 − 2x + x2 is a polynomial solution. 12 2! 2 1−1−3 1−1−3 2−1−3 2 1 3 3 1 • When λ = 3, y = 1+ x+ · x − x = 1−3x− x2 − x3 is a polynomial 2 2 2 1 1 2 3! 2 6 solution. 4. Consider the differential equation: x(x − 1)y 00 + 6x2 y 0 + 3y = 0, x > 0, (a) Show that x = 0 is a regular singular point. Find the indicial equation and the exponents at x = 0. Page 8 Answer. Since x(x − 1)y 00 + 6x2 y 0 + 3y = 0, then y 00 + Let p(x) = 3 6x2 y0 + y = 0. x(x − 1) x(x − 1) 6x2 6x 3 = and q(x) = , then x(x − 1) x−1 x(x − 1) 6x2 1 = −6x2 · x−1 1−x ∞ X = −6x2 xk xp(x) = k=0 = −6 = −6 ∞ X k=0 ∞ X xk+2 xk for all |x| < 1 k=2 1 3x = −3x · x−1 1−x ∞ X xk = −3x · x2 q(x) = k=0 = −3 = −3 ∞ X k=0 ∞ X xk+1 xk , for all |x| < 1. k=1 So x = 0 is a regular singular point, and p0 = q0 = 0. So the indicial equation is: 0 = r(r − 1) + p0 r + q0 = r(r − 1) = r2 − r. Then the exponents at x = 0 are r1 = 1 and r2 = 0. (b) Find the general series solutions near x = 0. What would you expect the radius of convergence of the series solution be? Answer. r Let φ(x, r) = x ∞ X k ak (r)x = k=0 ∞ X ak (r)xk+r for some {ak (r)}∞ k=0 which will be determined k=0 later, then φx (x, r) = ∞ X (k + r)ak (r)xk+r−1 k=0 φxx (x, r) = ∞ X (k + r)(k + r − 1)ak (r)xk+r−2 . k=0 Page 9 Then x(x − 1)φxx (x, r) + 6x2 φx (x, r) + 3φ(x, r) ∞ ∞ ∞ X X X = (x2 − x) (k + r)(k + r − 1)ak (r)xk+r−2 + 6x2 (k + r)ak (r)xk+r−1 + 3 ak (r)xk+r k=0 k=0 k=0 ∞ ∞ ∞ X X X k+r k+r−1 = (k + r)(k + r − 1)ak (r)x − (k + r)(k + r − 1)ak (r)x + 6(k + r)ak (r)xk+r+1 k=0 ∞ X + k=0 k=0 3ak (r)xk+r k=0 ∞ ∞ X X k−1+r = (k − 1 + r)(k − 1 + r − 1)ak−1 (r)x − (k + r)(k + r − 1)ak (r)xk+r−1 k=1 ∞ X + = k=0 6(k − 2 + r)ak−2 (r)xk−2+r+1 + k=2 ∞ X ∞ X 3ak−1 (r)xk−1+r k=1 (k + r − 1)(k + r − 2)ak−1 (r)xk+r−1 − k=1 ∞ X + ∞ X (k + r)(k + r − 1)ak (r)xk+r−1 k=0 6(k + r − 2)ak−2 (r)xk+r−1 + k=2 ∞ X 3ak−1 (r)xk+r−1 k=1 = r(r − 1)a0 (r)xr − r(r − 1)a0 (r)xr−1 − (1 + r)ra1 (r)xr ∞ X + [(k + r − 1)(k + r − 2)ak−1 (r) − (k + r)(k + r − 1)ak (r) + 6(k + r − 2)ak−2 (r) + 3ak−1 (r)] xk+r−1 + k=2 = −r(r − 1)a0 (r)xr−1 + [r(r − 1)a0 (r) − r(r + 1)a1 (r) + 3a0 (r)]xr ∞ X + {−(k + r)(k + r − 1)ak (r) + [(k + r − 1)(k + r − 2) + 3]ak−1 (r) + 6(k + r − 2)ak−2 (r)} xk+r−1 k=2 Define the sequence {ak (r)}∞ k=1 by the following recurrence relation: r(r − 1)a0 (r) − r(r + 1)a1 (r) + 3a0 (r) = 0, −(k + r)(k + r − 1)ak (r) + [(k + r − 1)(k + r − 2) + 3]ak−1 (r) + 6(k + r − 2)ak−2 (r) = 0, ∀k ≥ 2. That is, we have a1 (r) = ak (r) = r(r − 1) + 3 a0 (r) r(r − 1) [(k + r − 1)(k + r − 2) + 3]ak−1 (r) + 6(k + r − 2)ak−2 , (k + r)(k + r − 1) ∀k ≥ 2. Let a0 (r) = 1, it’s see that all ak (1)’s are well defined, let y1 (x) = φ(x, 1), then y1 (x) is a solution. Let ∂ a0 (r) = r, and y2 (x) = (rφ(x, r)) is another solution. The general solution is: ∂r r=0 y(x) = C1 y1 (x) + C2 y2 (x). Page 10 Since xp(x) = −6 ∞ X k 2 x and x q(x) = −3 ∞ X k=2 xk for all |x| < 1, then the radius of convergence of the k=1 general series solution would be at least 1. 9 5. Find the general series solution near x = 0 to x y + xy + x − 4 2 00 Since x2 y 00 + xy 0 + x2 − Answer. 9 4 0 2 y = 0. y = 0, then 1 0 x2 − 49 y + y + y = 0. x x2 00 x2 − 49 1 9 and q(x) = , then x = 0 is a singular point. Notice that xp(x) = 1 and x2 q(x) = x2 − , 2 x x 4 both are analytic at x = 0, then x = 0 is a regular singular point. So p0 = 1 and q0 = − 94 , which implies that the indicial equation is: 9 9 0 = r(r − 1) + p0 r + q0 = r(r − 1) + r − = r2 − . 4 4 Let p(x) = So the exponents at x = 0 are r1 = 3 2 and r2 = − 23 . Let ψ(x, r) := xr ∞ X ak (r)xk = k=0 ∞ X ak (r)xk+r for some k=0 differentiable functions ak (r), for all x > 0, then ψx (x, r) = ψxx (x, r) = ∞ X k=0 ∞ X (k + r)ak (r)xk+r−1 (k + r)(k + r − 1)ak (r)xk+r−2 . k=0 So we get 9 x ψxx (x, r) + xψx (x, r) + x − ψ(x, r) 4 ∞ ∞ ∞ X X 9 X 2 k+r−2 k+r−1 2 = x (k + r)(k + r − 1)ak (r)x +x (k + r)ak (r)x + x − ak (r)xk+r 4 2 = 2 k=0 ∞ X ∞ X k=0 k=0 (k + r)(k + r − 1)ak (r)xk+r + k=0 (k + r)ak (r)xk+r + ∞ X k=0 ak (r)xk+r+2 − k=0 ∞ X 9 4 ak (r)xk+r k=0 ∞ ∞ X X 9 = (k + r)(k + r − 1) + k + r − ak (r)xk+r + ak (r)xk+r+2 4 k=0 k=0 ∞ ∞ X X 9 = (k + r)2 − ak (r)xk+r + ak−2 (r)xk+r 4 k=2 k=0 ∞ X 9 9 9 2 r 2 1+r 2 = r − a0 (r)x + (1 + r) − a1 (r)x + (k + r) − ak (r) + ak−2 (r) xk+r . 4 4 4 k=2 Page 11 Let a1 (r) = 0, define the sequence {ak (r)}∞ k=1 by the recurrence relation: 9 (k + r)2 − ak (r) + ak−2 (r) = 0, for all k ≥ 2. 4 That is, ak (r) = −1 (k + r)2 − 9 4 · ak−2 (r), for all k ≥ 1. Then a2k+1 (r) = 0, and a2k (r) = (−1)k a0 (r) k Y 1 , 2− 9 (2i + r) 4 i=1 for all k ≥ 0. So we get ψ(x, r) = x r ∞ X a2k (r)x 2k r = a0 (r)x + a0 (r)x k=0 9 x2 ψxx (x, r) + xψx (x, r) + x2 − 4 r ∞ X k (−1) k=1 9 2 ψ(x, r) = r − a0 (r)xr . 4 k Y 1 (2i + r)2 − i=1 ! 9 4 x2k (2) (3) 3 3 2 9 2 . By For any 1 ≤ i ≤ k, then (2i + ) − = 4i + 3i > 0, which implies that we can define φ1 (x) = φ x, 2 4 2 (3), then φ1 (x) is a solution . Notice that ! ∞ k Y 3 3 X 1 k φ1 (x) = x 2 + x 2 x2k (−1) 9 3 2 ) − (2i + 2 4 i=1 k=1 " ! # ∞ k X Y 3 1 = x2 1 + (−1)k x2k . (2i) · (2i + 3) i=1 k=1 Let’s find another solution φ2 (x): Since r1 − r2 = 3, but a3 (r) = 0 which is well-defined. For any 2 ≤ i ≤ k, 3 2 9 we have 2i − − = 4i2 − 6i = i(2i − 3) ≥ 2 · (2 · 2 − 3) = 2 > 0. Let a0 (r) = 1, then 2 4 a2 (r) = (−1)a0 (r) · = 1 (2 · 1 + r)2 − −1 r+ r + 12 7 2 a2k (r) = (−1)k+1 · a2 (r) · k Y 1 (2i + r)2 − i=2 k = 9 4 9 4 Y (−1)k 1 · . 7 1 r + 2 r + 2 i=2 (2i + r)2 − 49 Page 12 3 Hence all a2k (r)’s are defined near r = Let φ2 (x) = φ x, − , since a0 (r) = 1, by (3), then φ2 (x) is a 2 solution which is linearly independent with φ1 (x). Notice that 3 φ2 (x) = φ x, − 2 " ! # ∞ k X Y 3 1 = x− 2 1 + x2k (−1)k 3 2 9 −4 i=1 2i − 2 k=1 " ! # ∞ k X Y 3 1 = x− 2 1 + x2k . (−1)k (2i)(2i − 3) − 32 . i=1 k=1 So the general solution is: y(x) = C1 φ1 (x) + C2 φ2 (x) ( " ∞ X 3 = C1 x 2 1 + (−1)k +C2 " x − 32 1+ 1 (2i) · (2i + 3) i=1 k=1 ( k Y ∞ X (−1)k k Y i=1 k=1 1 (2i)(2i − 3) 6. Consider the differential equation: x3 y 00 + αxy 0 + βy = 0, x > 0, where α and β are real constants and α 6= 0. (a) Show that x = 0 is an irregular singular point. Answer. Since x3 y 00 + αxy 0 + βy = 0, then y 00 + Let p(x) = α 0 β y + 3 y = 0. 2 x x α β and q(x) = 3 , then 2 x x xp(x) = α , x and x2 q(x) = So x = 0 is irregular. (b) Find a solution of the from ∞ X an xr+n with a0 = 1. n=0 Page 13 ! β . x #) x 2k ! #) x2k . Answer. Let φ(x, r) = ∞ X ak (r)xk+r with a0 (r) = 1, and {ak (r)}∞ k=1 will be determined later, then k=0 φx (x, r) = ∞ X (k + r)ak (r)xk+r−1 k=0 φxx (x, r) = ∞ X (k + r)(k + r − 1)ak (r)xk+r−2 . k=0 Then we have x3 φxx (x, r) + αxφx (x, r) + βφ(x, r) ∞ ∞ ∞ X X X k+r+1 k+r = (k + r)(k + r − 1)ak (r)x +α (k + r)ak (r)x +β ak (r)xk+r k=0 = ∞ X k=0 (k − 1 + r)(k − 1 + r − 1)ak−1 (r)x k+r k=0 +α k=1 = ∞ X ∞ X (k + r)ak (r)x k+r ∞ X +β k=0 (k + r − 1)(k + r − 2)ak−1 (r)xk+r + α k=1 ∞ X k=0 (k + r)ak (r)xk+r + β k=0 = αra0 (r)xr + βa0 (r)xr + ∞ X ak (r)xk+r ∞ X ak (r)xk+r k=0 [(k + r − 1)(k + r − 2)ak−1 (r) + α(k + r)ak (r) + βak (r)]xk+r k=1 ∞ X = (αr + β)xr + + [(k + r − 1)(k + r − 2)ak−1 (r) + (α(k + r) + β)ak (r)]xk+r . k=1 Then αr + β = 0. Let {ak (r)}∞ k=1 be defined by the following recurrence relation: (k + r − 1)(k + r − 2)ak−1 (r) + (α(k + r) + β)ak (r) = 0, for all k ≥ 1. That is, we get β r=− , α ak (r) = − and (k + r − 1)(k + r − 2) · ak−1 (r), α(k + r) + β for all k ≥ 1. β Since r = − , then α α(k + r) + β = αk + αr + β = αk 6= 0, for all k ≥ 1. β β So we know that φ x, − is well defined. In summary, we know that y(x) = φ x, − is a solution. α α Notice that ak (r) = k Y (−1)(i + r − 1)(i + r − 2) α(i + r) + β i=1 k = (−1) k Y (i + r − 1)(i + r − 2) i=1 αi + αr + β Page 14 , for all k ≥ 1. So we get β β k i − + 1 i − + 2 Y α α β = (−1)k ak − α αi i=1 k β (−1)k Y β i− +1 = i− +2 , α α αk · k! for all k ≥ 1. i=1 7. Solve the transport equation with the initial condition: ut + 3ux = 0, x ∈ R, t > 0. u(x, 0) = cos(x), x ∈ R. Answer. For each x ∈ R, since u(x, 0) = cos(x), then we need to find the characteristic curve through the point (0, x, cos(x)), that is, we should solve: 0 t (s) = 1 0 x (s) = 3 u0 (s) = 0, t(0) = 0, x(0) = x, u(0) = cos(x). So we get t(s) = s, x(s) = x + 3s, and u(s) = cos(x). Notice that u(s) = u(x(s), t(s)), then u(x + 3t, t) = cos(x). Since x ∈ R is arbitrary, then the solution is: u(x, t) = cos(x − 3t) for x ∈ R and t ≥ 0. 8. Solve the transport equation with the initial condition: ut + 2ux = ex+t , x ∈ R, t > 0. u(x, 0) = sin(x), x ∈ R. Answer. For each x ∈ R, since u(x, 0) = cos(x), then we need to find the characteristic curve through the point (0, x, cos(x)), that is, we should solve: 0 t 0(s) = 1 x (s) = 2 u0 (s) = ex(s)+t , t(0) = 0, x(0) = x, u(0) = sin(x). Page 15 So we get t(s) = s, x(s) = x + 2s, 1 and u(s) = [ex+3s − ex ] + sin(x). 3 Notice that u(s) = u(x(s), t(s)), then 1 u(x + 2t, t) = [ex+3t − ex ] + sin(x). 3 Since x ∈ R is arbitrary, then the solution is: 1 u(x, t) = [ex+t − ex−2t ] + sin(x − 2t) for x ∈ R and t ≥ 0. 3 9. Determine whether the method of separation of variables can be used to replace the given PDEs by a pair of ordinary differential equation. If so, find the equations. (a) xuxx + ut = 0. Answer. Let u(x, t) = X(x)T (t), then ux = X 0 (x)T (t), uxx = X 00 (x)T (t), and ut = X(x)T 0 (t). Then 0 = xuxx + ut = xX 00 (x)T (t) + X(x)T 0 (t). So we get − xX 00 (x) T 0 (t) = = λ. X(x) T (t) So we can use separation of variables, and the ODEs are: −xX 00 (x) − λX(x) = 0, and T 0 (t) − λT (t) = 0. (b) uxx + uxt + ut = 0. Answer. Let u(x, t) = X(x)T (t), then ux = X 0 (x)T (t), uxx = X 00 (x)T (t), ut = X(x)T 0 (t), and uxt = X 0 (x)T 0 (t) Then 0 = uxx + uxt + ut = X 00 (x)T (t) + X 0 (x)T 0 (t) + X(x)T 0 (t) So we get − T 0 (t) X 00 (x) = = λ. X 0 (x) + X(x) T (t) So we can use separation of variables, and the ODEs are: −xX 00 (x) − λ[X 0 (x) + X(x)] = 0, Page 16 and T 0 (t) − λT (t) = 0. (c) uxx + (x + y)uyy = 0. Answer. Let u(x, y) = X(x)Y (y), then ux = X 0 (x)Y (y), uxx = X 00 (x)Y (y), uy = X(x)Y 0 (y), and uyy = X(x)Y 00 (y). Then 0 = uxx + (x + y)uyy = X 00 (x)Y (y) + (x + y)X(x)Y 00 (y). So we get − X 00 (x) Y 00 (y) = . (x + y)X(x) Y (y) So we can not use separation of variables. (d) [p(x)ux ]x − r(x)utt = 0. Answer. Let u(x, t) = X(x)T (t), then ux = X 0 (x)T (t), uxx = X 00 (x)T (t), ut = X(x)T 0 (t), and utt = X(x)T 00 (t) Then 0 = [p(x)ux ]x − r(x)utt = [p(x)X 0 (x)T (t)]x − r(x)X(x)T 00 (t) = T (t)[p(x)X 0 (x)]x − r(x)X(x)T 00 (t). So we get T 00 (t) [p(x)X 0 (x)]x = = λ. r(x)X(x) T (t) So we can use separation of variables, and the ODEs are: [p(x)X 0 (x)]x − λr(x)X(x) = 0, and T 00 (t) − λT (t) = 0. (e) uxx + uyy + xu = 0. Answer. Let u(x, y) = X(x)Y (y), then ux = X 0 (x)Y (y), uxx = X 00 (x)Y (y), and uyy = X(x)Y 00 (y). Then 0 = uxx + uyy + xu = X 00 (x)Y (y) + X(x)Y 00 (y) + xX(x)Y (y) = [X 00 (x) + xX(x)]Y (y) + X(x)Y 00 (y). Page 17 So we get − X 00 (x) + xX(x) Y 00 (y) = = λ. X(x) Y (y) So we can use separation of variables, and the ODEs are: X 00 (x) + xX(x) + λX(x) = 0, and Y 00 (y) − λY (y) = 0. 10. Solve y 00 + y = 0, y(0) = 0 and y 0 (π) = 1 if the solution exists. Answer. Let y = erx to be a solution to y 00 + y = 0, then y 0 = rerx y 00 = r2 erx 0 = y 00 + y = r2 erx + erx = erx (r2 + 1). So r2 + 1 = 0, that is, r = ±i. Then the general solution to y 00 + y = 0 is: y = C1 cos(x) + C2 sin(x). Since y(0) = 0, that is, C1 = 0. Since y 0 (x) = −C1 sin(x) + C2 cos(x) and y 0 (π) = 1, that is, C2 cos(π) = 1. So C2 = −1. Then the solution is: y = − sin(x). 11. Solve y 00 + y = 0, y 0 (0) = 1 and y(π) = 0 if the solution exists. Answer. Let y = erx to be a solution to y 00 + y = 0, then y 0 = rerx y 00 = r2 erx 0 = y 00 + y = r2 erx + erx = erx (r2 + 1). So r2 + 1 = 0, that is, r = ±i. Then the general solution to y 00 + y = 0 is: y = C1 cos(x) + C2 sin(x). Since y(π) = 0, that is, C1 = 0. Since y 0 (x) = −C1 sin(x) + C2 cos(x) and y 0 (0) = 1, that is, C2 cos(0) = 1. So C2 = 1. Then the solution is: y = sin(x). Page 18 12. Use separation of variables to find PDE: BC: IC: Answer. the solution to the heat conduction problem: 100uxx = ut , 0 < x < 1, t > 0, u(0, t) = 0, u(1, t) = 0, t > 0, u(x, 0) = sin(2πx) − sin(5πx), 0 ≤ x ≤ 1. Let’s first look at the PDE and the homogeneous boundary condition: ( ut = 100uxx , 0 < x < 1, t > 0, BC: u(0, t) = 0, u(1, t) = 0, t > 0. (4) A separated solution to (4) is of the form u(x, t) = X(x)T (t), then ut (x, t) = X(x)T 0 (t) ux (x, t) = X 0 (x)T (t) uxx (x, t) = X 00 (x)T (t) 0 = ut − 100uxx = X(x)T 0 (t) − 100X 00 (x)T (t). So we get − X 00 (x) T 0 (t) =− := λ. 100T (t) X(x) (5) X 00 (x) , then λt (x, t) = 0. By (5), then λx (x, t) = 0. So λ(x, t) is a constant, which is called X(x) the separation constant. Since u(0, t) = u(1, t) = 0, then X(x) satisfies ( 00 X + λX = 0, (6) X(0) = X(1) = 0. Let λ(x, t) = − We are expected to find non-zero solutions to (6). Let X = erx be a solution to X 00 + λX = 0, then X 0 = rerx , X 00 = r2 erx , and 0 = X 00 + λX = r2 erx + λerx = erx (r2 + λ). Then r2 + λ = 0. We have three cases: a. If λ = 0, that is, r2 = 0. So r = 0, which implies that the general solution to X 00 + λX = X 00 = 0 is X(x) = C1 + C2 x. Since X(0) = X(1) = 0, then C1 = C1 + C2 L = 0. Then C1 = C2 = 0. Page 19 b. If λ > 0, that is, λ = µ2 for some µ > 0. So r2 + µ2 = 0, that is, r = ±µi, which implies that the general solution to X 00 + λX = X 00 + µ2 X = 0 is X(x) = C1 cos(µx) + C2 sin(µx). Since X(0) = X(1) = 0, then C1 = C1 cos(µ) + C2 sin(µ) = 0. Then C1 = 0, and C2 sin(µ) = 0. In order to make X to be a non-zero solution, then we need sin(µ) = 0, that is, µ = nπ for some n ∈ N, that is, λ = µ2 = n2 π 2 . For λ = n2 π 2 , then all solution to (6) are Xn (x) = C sin(nπx). c. If λ < 0, that is, λ = −µ2 for some µ > 0. So r2 − µ2 = 0, that is, r = ±µ, which implies that the general solution to X 00 + λX = X 00 − µ2 X = 0 is X(x) = C1 eµx + C2 e−µx . Since X(0) = X(1) = 0, then C1 + C2 = C1 eµ + C2 e−µ = 0. Since µ > 0, then C1 = C2 = 0. In summary, (6) has non-zero solution if and only if λ = n2 π 2 for some n ∈ N. When λ = n2 π 2 in (6), all non-zero solutions to (6) are X(x) = C sin(nπ). When λ = n2 π 2 , by (5), then 0 = T 0 (t) + 100λT (t) = T 0 (t) + 100n2 π 2 T (t) (7) It’s easy to see that all solutions to (7) are given by: 2 π2 t T (t) = Ce−100n . In summary, for any n ∈ N, we can find a non-zero solution to (4): 2 π2 t un (x, t) = e−100n sin(nπx). Since u(x, 0) = sin(2πx) − sin(5πx), then the solution to the initial value problem is given by: 2 2 u(x, t) = e−400π t sin(2πx) − e−2500π t sin(5πx). 13. Use separation of variables PDE: BC: IC: to find the solution to the heat conduction problem: uxx = 4ut + u, 0 < x < 2, t > 0, u(0, t) = 0, u(2, t) = 0, t > 0, πx u(x, 0) = 2 sin − sin(πx) + 4 sin(2πx), 2 Page 20 0 ≤ x ≤ 2. Answer. Let’s first look at the PDE and the homogeneous boundary condition: ( uxx = 4ut + u, 0 < x < 2, t > 0, BC: u(0, t) = 0, u(2, t) = 0, t > 0. (8) A separated solution to (8) is of the form u(x, t) = X(x)T (t), then ut (x, t) = X(x)T 0 (t) ux (x, t) = X 0 (x)T (t) uxx (x, t) = X 00 (x)T (t) 0 = 4ut + u − uxx = 4X(x)T 0 (t) + X(x)T (t) − X 00 (x)T (t) = 4X(x)T 0 (t) + [X(x) − X 00 (x)]T (t). So we get [X 00 (x) − X(x)]T (t) = 4X(x)T 0 (t), that is, − X 00 (x) − X(x) 4T 0 (t) =− := λ. T (t) X(x) (9) X 00 (x) − X(x) , then λt (x, t) = 0. By (9), then λx (x, t) = 0. So λ(x, t) is a constant, which is X(x) called the separation constant. Since u(0, t) = u(2, t) = 0, then X(x) satisfies ( 00 X + (λ − 1)X = 0, (10) X(0) = X(2) = 0. Let λ(x, t) = − We are expected to find non-zero solutions to (10). Let X = erx be a solution to X 00 +λX = 0, then X 0 = rerx , X 00 = r2 erx , and 0 = X 00 + (λ − 1)X = r2 erx + (λ − 1)erx = erx (r2 + λ − 1). Then r2 + λ − 1 = 0. We have three cases: a. If λ − 1 = 0, that is, r2 = 0. So r = 0, which implies that the general solution to X 00 + (λ − 1)X = X 00 = 0 is X(x) = C1 + C2 x. Since X(0) = X(2) = 0, then C1 = C1 + C2 2 = 0. Then C1 = C2 = 0. b. If λ − 1 > 0, that is, λ − 1 = µ2 for some µ > 0. So r2 + µ2 = 0, that is, r = ±µi, which implies that the general solution to X 00 + (λ − 1)X = X 00 + µ2 X = 0 is X(x) = C1 cos(µx) + C2 sin(µx). Since X(0) = X(2) = 0, then C1 = C1 cos(2µ) + C2 sin(2µ) = 0. Page 21 Then C1 = 0, and C2 sin(2µ) = 0. In order to make X to be a non-zero solution, then we need sin(2µ) = 0, that is, 2µ = nπ for some n ∈ N, nπ 2 nπ 2 that is, λ = µ2 − 1 = − 1. For λ = − 1, then all solution to (10) are 2 2 nπx Xn (x) = C sin . 2 c. If λ − 1 < 0, that is, λ − 1 = −µ2 for some µ > 0. So r2 − µ2 = 0, that is, r = ±µ, which implies that the general solution to X 00 + (λ + 1)X = X 00 − µ2 X = 0 is X(x) = C1 eµx + C2 e−µx . Since X(0) = X(2) = 0, then C1 + C2 = C1 e2µ + C2 e−2µ = 0. Since µ > 0, then C1 = C2 = 0. nπ 2 nπ 2 In summary, (10) has non-zero solution if and only if λ = − 1 for some n ∈ N. When λ = −1 2 nπ 2 x . in (10), all non-zero solutions to (10) are X(x) = C sin 2 nπ 2 When λ = − 1, by (9), then 2 0 = 4T 0 (t) + λT (t) nπ 2 0 − 1 T (t) = 4T (t) + 2 (11) It’s easy to see that all solutions to (11) are given by: − 14 h T (t) = Ce 2 ( nπ 2 ) i −1 t . In summary, for any n ∈ N, we can find a non-zero solution to (8): − 41 un (x, t) = e Since u(x, 0) = 2 sin πx 2 h 2 ( nπ 2 ) i −1 t sin nπ 2 . − sin(πx) + 4 sin(2πx), then the solution to the initial value problem is given by: u(x, t) = 2e − 14 h ( π2 ) 2 i −1 t sin πx 2 1 − e− 4 [π Page 22 ]t sin(πx) + 4e− 14 [4π2 −1]t sin(2πx). 2 −1