HOMEWORK 1 SOLUTIONS

Section 1.1 : #2, 3, 11, 12 #2) Which of the following operators are linear? (This is accomplished by checking if Lu = cLu and L(u + v ) = Lu + LV are satised). (a) Lu = ux + xuy (b) Lu = ux + uuy (c) Lu = ux + u2 y (d) Lu = ux + uy + 1 (e) Lu = 1 + x2 cos(y )ux + uyxy arctan( x y )u Answer: Only (a) and (e) are Linear. (b), (c), and (d) do not satisfy L(cu) = cL(u). #3) For each of the equations below state the order and type. Equation (a) (b) (c) (d) (e) (f ) (g ) (h) Order 2 2 3 2 2 1 1 4 Type Linear Inhomogeneous Linear Homogeneous Non-Linear Linear Inhomogeneous Linear Homogeneous Non-Linear Linear Homogeneous Non-Linear

#11 Verify that u(x, y ) = f (x)g (y ) is a solution for all pairs of dierentiable functions f and g of one variable to the equation: uuxy = ux uy Answer: ux = f (x)g (y ) uy = f (x)g (y ) uxy = f (x)g (y ) Then we have: uuxy = f (x)g (y )f (x)g (y ) = f (x)g (y )f (x)g (y ) = ux uy #12 Verify by substitution that: un (x, y ) = sin(nx) sinh(ny ) is a solution of uxx + uyy = 0 for every n > 0. Answer: (un )xx = n2 sin(nx) sinh(ny ) and (un )yy = n2 sin(nx) sinh(ny ). When we add these together, we get (un )xx + (un )yy = 0.
1

Section 1.2 : #1, 3, 7, 8 #1 Solve the rst order equation with initial condition: 2ut + 3ux = 0 u(x, 0) = sin(x) Answer: Following the Geometric or Coordinate Method, we get: u(x, t) = f (2x 3t) We solve for f (x) using the initial condition. u(x, 0) = f (2x) = sin(x) Then f (x) = sin( x 2 ), which means 3 u(x, t) = sin(x t) 2 #3 Solve (1 + x2 )ux + uy = 0 and sketch some characteristic curves. Answer: Following the Geometric Method : dy 1 = dx 1 + x2 y = arctan(x) + c c = y arctan(x) u(x, y ) = f (c) = f (y arctan(x)) The characteristic curves are graphs of the form y = arctan(x) + c for any value of c. #7 Solve aux + buy + cu = 0. Answer: Following the Coordinate Method : Let x = ax + by and y = bx ay . Then we have: aux + buy = (a2 + b2 )ux (a2 + b2 )ux + cu = 0 c ux = 2 u a + b2 c ln(u) = 2 x + f (y ) a + b2
cx

u(x , y ) = f (y )e a2 +b2 u(x, y ) = f (bx ay )e

c(ax+by ) a 2 +b 2

Alternative Solution: Rewrite the equation as: ux uy a +b +c=0 u u

Then let v = ln u and solve: avx + bvy + c = 0 This has the homogenous solution vh = f (bx ay ) and a particular solution vp =
cx a .

Then v (x, y ) = f (bx ay ) + u(x, y ) = f (bx ay )e cx a

cx a

Note: This solution does not appear valid for a = 0, but with an appropriate choice of f (bx ay ) it is equivalent to the solution reached using the CoordinateM ethod. #8 Solve ux + uy + u = ex+2y with u(x, 0) = 0. Answer: One possible approach includes using the solution to the previous problem to nd the solution to the homogenous equation and then nding a particular solution. Heres another approach using a change of coordinates and an integrating factor. Let x = x + y and y = x y . The equation to solve becomes: 2ux + u = e 2 1 1 3x y ux + u = e 2 2 2 Multiply the equation by the integrating factor e 2 to get: x 1 x 1 4x y e 2 ux + ue 2 = e 2 2 2 x 1 4x y = e 2 e2u 2 x x 1 4x y e 2 dx e 2 u dx = 2 x x 1 4x y e 2 u = e 2 + f (y ) 4 x 1 3x y u(x , y ) = e 2 + e 2 f (y ) 4 xy 1 u(x, y ) = ex+2y + e 2 f (x y ) 4 Using the initial condition to solve for f , we have: x 1 0 = u(x, 0) = ex + e 2 f (x) 4 Then f (x) =
x 1 3 2 4 e , x 3x y

and we get: 1 x+2y 1 x2y e e 4 4

u(x, y ) = .

Section 1.3 : #7, 8 #7 Derive the equation for heat ow in a ball where the temperature depends only on the spherical coordinate r = cut = (u) In this problem, we are told that c, , and are constants, so ut = C u = C (uxx + uyy + uzz ) Using the chain rule, we rewrite uxx in terms of derivatives in terms of r.
2 uxx = (ur rx )x = urx rx + ur rxx = urr rx + ur rxx

x2 + y 2 + z 2 .

Answer: We start with the general heat equation:

Also, r= rx = rxx Then x2 + y 2 + z 2 x r 1 x2 = 3 r r

uxx = urr The expressions for uyy

x2 1 x2 + ur 3 2 r r r and uzz are similar. When added together, we get 2ur r

1 |x|3 +1

ut = C urr +

#10 Given f (x) is continuous and |f (x)| ball of radius R. Then f d x = lim R all space lim = lim = lim Similarly, f d x lim all space Then by the squeeze theorem:
R R

for all x. Let B (R) be a f ( x) n dS

B (R)

f (x)d x = lim
B (R )

|f (x) n |dS = lim

B (R)

|f (x)|| n |dS
B (R)

|f (x)|dS lim
B (R)

B (R)

1 4R dS = lim 3 =0 R R + 1 R3 + 1 |f (x) n |dS 0

B (R)

B (R) 2

|x|3

1 dS +1

f d x =0 all space

HOMEWORK 2 SOLUTIONS
Section 1.5 : #1, 2, 3, 4(a) (b) #1) Solve the boundary value problem: uxx + u = 0 u(0) = 0 and u(L) = 0. Answer: The solution to the ODE is: u(x) = A sin(Bx + C ) If A = 0, then the solution is u(x) 0. If A = 0, then we can solve for the variables by substituting the general solution into the ODE. uxx + u = (A AB 2 ) sin(Bx + C ) = 0 A(1 B 2 ) = 0 B = 1 Using the boundary conditions we have: u(0) = A sin(C ) = 0 Then C = n with n an integer. Also, u(L) = A sin(BL + C ) = 0 BL + C = m L + n = m L = m So, if L = m for some integer m, then u(x) = A sin(x + n ) is a solution. If L = m , then u(x) 0 is the only solution. #2) Consider the problem u (x) + u (x) = f (x) 1 [u (l) + u(l)] 2 (a) Does the following problem have a unique solution? (b) Does a solution necessarily exist? Answer: (a) The problem does not have a unique solution. u (0) = u(0) = Proof. Let u1 and u2 be solutions to the problem. Let v (x) = u1 (x) u2 (x). Then v (x) is a solution of u (x) + u (x) = 0
1

u (0) = u(0) = Then v (x) = A + Bex , and v (0) = v (0) =

1 [u (l) + u(l)] 2

1 [v (l) + v (l)] 2 1 B = A + B = [Bel + A + Bel ] 2 A B = A + B = 2 A = 2B

v (x) = B (ex 2), for any real number B Then if u(x) is a solution of the initial problem, u2 (x) = u1 (x) + ex 2 is also a solution. (b)A solution does not necessarily exist unless f (x) satises the condition
l

f (x)dx = 0
0

Proof. We integrate both sides of the ODE and use the conditions on u at x = 0 and at x = l.
l l

[u (x) + u (x)]dx =
0 0 l

f (x)dx f (x)dx
0 l

u (l) + u(l) u (0) u(0) = 2u(0) u(0) u(0) =

0 l

f (x)dx f (x)dx
0

0=

#3 Solve the boundary value problem u (x) = 0 if 0 < x < 1, u (0) + ku(0) = 0 u (1) ku(1) = 0 Answer: The solution to the ODE on the interval 0 < x < 1 is: u(x) = Ax + B Using the boundary conditions, we have: A = Bk Bk k (Bk + B ) = 0

Case 1 (+k ): The equations from the boundary conditions are A = Bk Bk 2 = 0 Then either B = 0 or k = 0, which means that A = 0. Then u(x) = B with B = 0 if k = 0. Case 2 (k ): The equations from the boundary conditions are A = Bk Bk (k 2) = 0 Then either B = 0, k = 0, or k = 2. Again, if B = 0 or k = 0, then A = 0. If k = 2, then A = 2B B if k = 0 and B R, 0 if k = 0 and k = 2, u(x) = 2Bx + B if k = 2 . The only degree 1 solution occurs when k = 2 in the k case. #4 Consider the Neumann problem: u = f (x, y, z ) in D, n u = 0 on boundary of D. (a) What can we add to any solution to get another solution? (b) Show that f (x, y, z )dxdydz = 0. D Answer: (a) We can add any solution to the homogeneous Neumann problem to any solution of the above Neumann problem. This include constants and ax + by + cz with a, b, c n = 0. (b) Using the divergence theorem and the information given in the PDE, we have: f (x, y, z )dxdydz =
D D

udxdydz udxdydz
D

= =
D

n udS n udS
D

= =
D

0dS = 0

Section 1.6 : #1, 4 #1 What are the types of the following equations? (a) uxx uxy + 2uy + uyy 3uyx + 4u = 0 Answer: Remember that uxy = uyx , then the equation becomes: uxx 4uxy + uyy + 2uy + 4u = 0 4 (1)(1) = 3 > 0 2 Then the equation is Hyperbolic. D= (b) 9uxx + 6uxy + uyy + ux = 0 2 Answer: D = 6 (9)(1) = 9 9 = 0. Then the equation is Parabolic. 2 #4 What is the type of the equation uxx 4uxy + 4uyy = 0? Answer: D = (2)2 (1)(4) = 0. It is Parabolic. Show by direct substitution that u(x, y ) = f (y + 2x) + xg (y + 2x) is a solution for arbitrary functions f and g . Answer: Compute the partial derivatives and then substitute. ux = 2f (y + 2x) + g (y + 2x) + 2xg (y + 2x) uxx = 4f (y + 2x) + 2g (y + 2x) + 2g (y + 2x) + 4xg (y + 2x) = 4f (y + 2x) + 4g (y + 2x) + 4xg (y + 2x) uxy = 2f (y + 2x) + g (y + 2x) + 2xg (y + 2x) uy = f (y + 2x) + xg (y + 2x) uyy = f (y + 2x) + xg (y + 2x) Then substituting, we have: uxx 4uxy + 4uyy = 4f (y + 2x) + 2g (y + 2x) + 2g (y + 2x) + 4xg (y + 2x) 8f (y + 2x) 4g (y + 2x) 8xg (y + 2x) + 4f (y + 2x) + 4xg (y + 2x) =0
2

Section 2.1 : #8, 9 #8 Consider the spherical wave equation: 2 utt = c2 urr + ur r (a) Change variables by making the substitution v = ru. Answer: First solve for u, then take partial derivatives. 1 v r 1 utt = vtt r 2 1 2 urr = vrr 2 vr + 3 v r r r u= Substituting into the spherical wave equation, we get 1 1 vtt = c2 vrr r r vtt = c2 vrr (b) Solve for v using the fact that utt = c2 uxx has a solution of the form u(x, t) = f (x + ct) + g (x ct) Answer: All that has changed here are the variables, so v (r, t) = f (r + ct) + g (r ct) is a solution of vtt = c2 vrr , which means u(r, t) = 1 1 f (r + ct) + g (r ct) r r

is a solution of the spherical wave equation. (c) Solve the spherical wave equation with initial conditions u(r, 0) = (r) ut (r, 0) = (r) with both and even functions of r. Answer: First, rewrite the initial conditions in terms of v , and then solve using the solution to the initial value problem wave equation given by (8) in the book. v (r, 0) = r(r) = L(r) vt (r, 0) = r (r) = M (r)

Then from the solution to the IVP, we have v (r, t) = v (r, t) = u(r, t) = 1 1 [L(r + ct) + L(r ct)] + 2 2c
r +ct

M (s)ds
r ct r +ct

1 1 [(r + ct)(r + ct) + (r ct)(r ct)] + 2 2c

s (s)ds
r ct r +ct

1 1 [(r + ct)(r + ct) + (r ct)(r ct)] + 2r 2rc

s (s)ds
r ct

#9 Solve the initial value problem uxx 3uxt 4utt = 0, u(x, 0) = x2 , ut (x, 0) = ex . Answer: By factoring the operator acting on u, we have 1 4(t x )(t + x )u = 0. 4 Noticing the similarities of this to the wave equation, which factors like (t cx )(t + cx )u = 0, you may guess that the solution were looking for has the form 1 u(x, t) = f (x t) + g (x + t). 4 One way to show this is to let v = (x + t )u. Then we have ut + ux = v 1 vt vx = 0 4 Then ux + ut = v (x, t) = h(x + 1 4 t). 1 ux + ut = h(x + 4 t) has a homogeneous solution of the form uh = f (x t). 4 It also has a particular solution up = g (x + 1 4 t) with g (s) = 5 h(s). Then 1 u(x, t) = f (x t) + g (x + t) 4 1 1 ut (x, t) = f (x t) + g (x + t) 4 4 Considering the initial conditions, we have u(x, 0) = f (x) + g (x) = x2 1 ut (x, 0) = f (x) + g (x) = ex 4

Solving for f and g , we integrate ex dx = 1 f (x) + g (x)dx 4 1 ex = f (x) + g (x) + c 4 1 ex = g (x) x2 + g (x) + c 4

So, we have 4 x e + x2 + c 5 4 x 1 2 e + x c f (x) = 5 5 1 4 x+ 1 t e 4 ext + x2 + t2 u(x, t) = 5 4 g (x) =

HOMEWORK 3 SOLUTIONS
Section 2.1 : #2, 7 #2) Solve utt = c2 uxx u(x, 0) = log(1 + x2 ) ut (x, 0) = 4 + x Answer: Using the solution to the initial value problem wave equation, we have u(x, t) = 1 1 log(1 + [x + ct]2 ) + log(1 + [x ct]2 ) + 2 2c
x+ct

4 + sds
xct

After combining the logarithms and evaluating the integral u(x, t) = 1 log (1 + [x + ct]2 )(1 + [x ct]2 ) + 4t + xt 2

#7) If both and are odd functions of x, show that solution u(x, t) of the wave equation is also an odd function of x for all t. Answer: Show u(x, t) = u(x, t). u(x, t) = u(x, t) = = 1 1 [(x + ct) + (x ct)] + 2 2c
x+ct

(s)ds
xct x+ct

1 1 [(x + ct) + (x ct)] + 2 2c

(s)ds
xct xct

1 1 [(x ct) (x + ct)] (s)ds 2 2c x+ct (let s = r, then ds = dr and we change the limits of integration) u(x, t) = = 1 1 [(x ct) (x + ct)] 2 2c
x+ct

(r)(dr)
xct x+ct

1 1 [(x ct) + (x + ct)] 2 2c = u(x, t)

(r)dr
xct

Section 2.2 : #1, 2, 3 #1) Use energy conservation of the wave equation to prove that the only solution with 0 and 0 is u 0.
1

Answer: We know u(x, 0) 0 and ut (x, 0) 0. 1 2 1 E (x, 0) = 2 1 = 2 1 E (x, 0) = 2 E (x, t) =

(ut (x, t))2 + T (ux (x, t))2 dx

(ut (x, 0))2 + T (x u(x, 0))2 dx

0)2 + T (x (0))2 dx

0dx = 0

Then E (x, t) 0, since t E = 0. 0= 1 2

(ut (x, t))2 + T (ux (x, t))2 dx(ut (x, t))2 + T (ux (x, t))2

Then by the vanishing theorem (ut (x, t))2 + T (ux (x, t))2 0 Then ut (x, t) 0 ux (x, t) 0 Then u(x, t) is a constant. However, u(x, 0) 0, then u(x, t) 0. #2) For a solution u(x, t) of the wave equation with c = 1, the energy density e, and the momentum density p are dened as 1 2 u + u2 x 2 t p = ut ux e= (a) Show that t e = x p and t p = x e. (b) Show that both e(x, t) and p(x, t) satisfy the wave equation. Answer: (a) We take the partial derivatives using the chain rule and use the fact that utt = uxx . t e = 1 2 t u2 t + ux 2 1 = (2ut utt + 2ux uxt ) 2 = ut utt + ux uxt

= ut uxx + ux utx = x (ut ux ) = x p

Do the same for the next part t p = t (ut ux ) = utt ux + ut uxt = uxx ux + ut uxt 1 1 = (u2 )x + (u2 )t 2 x 2 t 1 2 = u + u2 t x = x e 2 x (b) Here are two methods to prove this, the second being the fastest. To show p satises the wave equation we take the second partial derivatives with respect to t and x, and then show that they are equal to each other using utt = uxx . ptt = (ut ux )tt = (utt ux + ut uxt )t = uttt ux + 2utt uxt + ut uxtt = (ut t)t ux + 2uxx utx + ut (utt )x = uxxt ux + 2uxx utx + ut uxxx = utxx ux + 2uxx utx + ut uxxx = (utx ux + ut uxx )x = (ut ux )xx = pxx To show e satises the wave equation we make use of part (a). ett = (et )t = (px )t = pxt = ptx = (pt )x = (ex )x = exx #3) Show that the wave equation has the following invariance properties. (a) Any translate u(x y, t), where y is xed is also a solution. Answer: We know utt (x, t) = c2 uxx (x, t) for any pair (x, t). 2 u(x y, t) = [ux (x y, t)x (x y )] x2 x = [ux (x y, t) 1] x = uxx (x y, t)x (x y ) = uxx (x y, t) 2 u(x y, t) = [ut (x y, t)t (t)] 2 t t = utt (x y, t) Then 2 u(x y, t) 2 u(x y, t) = utt (x y, t) = c2 uxx (x y, t) = c2 2 t x2

(b) Any derivative, say ux , is also a solution. Answer: We can change the order of partial derivatives, so 2 ux (x, t) = uxtt (x, t) t2 = uttx (x, t) = (utt (x, t))x = (c2 uxx (x, t))x = c2 2 ux (x, t) x2

(c) The dilated function u(ax, ay ) is also a solution. Answer: Use the chain rule and utt (ax, at) = c2 uxx (ax, at). 2 u(ax, at) = a2 utt (ax, at) t2 = a2 c2 uxx (ax, at) = c2 2 u(ax, at) x2

Section 2.3 : #2, 4 #2) Consider a solution of the diusion equation in {0 x , 0 t < }. (a) Let M (T ) be the maximum of the solution in the rectangle {0 x , 0 t T }. Does M (T ) increase or decrease as a function of time? Answer: M (T ) is increasing as a function of time. By the maximum principle the maximum occurs on the bottom or lateral sides of the rectangle. Suppose it occurs on the lateral side for some time t = S > T . Then M (T ) < M (S ). So, M (T ) cannot be decreasing. (b)) Let m(T ) be the minimum of the solution in the rectangle {0 x , 0 t T }. Does m(T ) increase or decrease as a function of time? Answer: m(T ) is decreasing as a function of time. By the maximum principle the minimum occurs on the bottom or lateral sides of the rectangle. Suppose it occurs on the lateral side for some time t = S > T . Then m(T ) > m(S ). So, m(T ) cannot be increasing.

#4) Consider the diusion equation ut = uxx in {0 < x < 1, 0 < t < } u(0, t) = u(1, t) = 0 u(x, 0) = 4x(1 x) (a) Show that for all (x, t) inside the rectangle 0 < u(x, t) < 1. Answer: From the strong maximum principle we know that the maximum and minimum of the diusion equation only occur on the boundary of the rectangle. On the lateral sides u 0. On the bottom u (x, 0) = 4 8x. So x =
1 2

is the only critical point, and 1 u( , 0) = 1. 2

Then the maximum of u(x, t) is 1 and the minimum is 0. Then by the strong maximum principle for all (x, t) inside the rectangle 0 < u(x, t) < 1. (b) Show that u(x, t) = u(1 x, t) for all t 0 and 0 x 1. Answer: We know that the solution to the Dirichlet problem for the diusion equation is unique, so we need only check that u(1 x, t) is a solution. Let v (x, t) = u(1 x, t). v (0, t) = u(1, t) = 0 = u(0, t) = v (1, t) v (x, 0) = u(1 x, 0) = 4(1 x)(1 (1 x)) = 4(1 x)x vt (x, t) = ut (1 x, t) vx (x, t) = ux (1 x, t) vxx (x, t) = uxx (1 x, t) = ut (1 x, t) = vt (x, t) Then u(1 x, t) is a solution of the Dirichlet problem. Then u(1 x, t) u(x, t) on the rectangle. (c) Use the energy method to show that function of t.
1 0

u2 dx is a strictly decreasing

Answer: Using the energy method and u(1, t) = u(0, t) = 0. 0 = ut uxx = (ut uxx )u = uut = uuxx 1 0 = (u2 )t (uux )x + (ux )2 2 1 1 1 2 (uux )x dx + (u )t dx 0= 2 0 0 = 0= 1 2 t 1 2 t
1 1

(ux )2 dx
0

u2 dx uux |1 0 dx +
0 1 1 0

(ux )2 dx

u2 dx +
0 0

(ux )2 dx

Then we have t However, if

1 1 1

u2 dx 2
0 0

(ux )2 dx 0

(ux )2 dx = 0
0

Then ux 0 by the vanishing theorem. Then u(x, t) = f (t), but we know that u(x, 0) = 4x(1 x). Then
1

(ux )2 dx > 0
0

which means that t Then

1 0 0

u2 dx < 0

u2 dx is strictly decreasing.

HOMEWORK 4 SOLUTIONS
Section 2.4 : #2, 4, 9, 11, 15 #2) Solve the diusion equation with the given initial condition: ut = kuxx in { < x < , 0 < t < } u(x, 0) = (x) = 1 3 if x > 0, if x < 0.

Solution: This problem is the same as the problem at the beginning of section 2.4. Then the solution is given by equation (8). You can substitute in for (y ), and then simplify the integral. However, there is an easier method for this particular problem. Consider 1 Q(x, t) = (u(x, t) 3). 2 This satises the diusion equation and satises the special initial condition: 1 if x > 0, Q(x, 0) = 0 if x < 0. Then, following the method on pages 46 47, yields
4kt 2 1 1 ep dp Q(x, t) = + 2 0 x 1 1 = + E rf 2 2 4kt x u(x, t) = 2 E rf . 4kt x

#4) Solve the diusion equation with the given initial condition: ut = kuxx in { < x < , 0 < t < } u(x, 0) = (x) = ex 0 if x > 0, if x < 0.

Solution: Again, this problem is the same as the problem at the beginning of section 2.4. Then the solution is given by equation (8). You can substitute in for (y ), and then simplify the integral. u(x, t) = 1 4kt 1 = 4kt

e
0

(xy )2 4kt

ey dy
y

e
0

(xy )2 4kt

dy

By completing the square, we rewrite the exponent as (y + 2kt x)2 + kt x, 4xt

1

ktx) and let p = (y+2 . Then dy = 4ktdp. After substituting and remem4kt bering to change the limits of integration, we have 1 u(x, t) = ektx 1 = ektx

2 ktx 4kt

ep dp
0

ep dp
0

2 ktx 4kt

ep dp

1 2kt x = ektx E rf 2 2 4kt 1 ktx 2kt x = e 1 E rf 2 4kt

#9) Solve the diusion equation with initial condition: ut = kuxx u(x, 0) = x2 . After dierentiating both sides with respect to x three times, we have (uxxx )t = k (uxxx )xx uxxx (x, 0) = 0. Then uxxx (x, t) = 0 is a solution to the above equation, but by uniqueness of solutions, we have: uxxx (x, t) 0. Integrating both sides with respect to x three times, yields u(x, t) = A(t)x2 + B (t)x + C (t). Next, use the initial condition and the diusion equation to determine the functions A(t), B (t), and C (t). u(x, 0) = A(0)x2 + B (0)x + C (0) = x2 Then A(0) = 1, B (0) = 0, and C (0) = 0. Furthermore, kuxx = ut k (2A(t)) = A (t)x2 + B (t)x + C (t) A (t) = 0 A(t) = constant B (t) = 0 B (t) = constant C (t) = k (2A(t))

Since A(t) is a constant, and A(0) = 1, we know A(t) = 1. Similarly B (t) = 0. Then we have C (t) = 2k . Since C (0) = 0, we know that C (t) = 2kt. Then the solution is given by u(x, t) = x2 + 2kt #11) Consider the diusion equation on the whole line with u(x, 0) = (x) (1) If (x) is an odd function of x, show that the solution u(x, t) is also. Proof. Let w(x, t) = u(x, t) + u(x, t). Then w(x, t) satises the diffusion equation, and w(x, 0) = u(x, 0) + u(x, 0) = (x) + (x) = (x) + (x) = 0. Then we have the initial condition diusion equation wt = kwxx w(x, 0) = 0 By uniqueness, w(x, t) 0.Then u(x, t) + u(x, t) = 0 u(x, t) = u(x, t). Thus, u(x, t) is an odd function of x. (2) If (x) is an even function of x, show that the solution u(x, t) is also. Proof. Let w(x, t) = u(x, t) u(x, t). Then w(x, t) satises the diffusion equation, and w(x, 0) = u(x, 0) u(x, 0) = (x) (x) = (x) (x) = 0. Then we have the initial condition diusion equation wt = kwxx w(x, 0) = 0 By uniqueness, w(x, t) 0.Then u(x, t) u(x, t) = 0 u(x, t) = u(x, t). Thus, u(x, t) is an even function of x. (3) Show that the analogous statements are true for the wave equation. Proof. The analogous statement for the wave equation involves specifying the oddness/evenness of both (x) and (x). Otherwise, the proof is the same.

If (x) and (x) are odd functions of x, show that u(x, t) is an odd function of x. Let w(x, t) = u(x, t) + u(x, t). Then w(x, t) satises the wave equation, and we have the initial condition wave equation wtt = c2 wxx , w(x, 0) = 0, wt (x, 0) = 0. By uniqueness, w(x, t) 0.Then u(x, t) + u(x, t) = 0 u(x, t) = u(x, t). Thus, u(x, t) is an odd function of x. If (x) and (x) are even functions of x, show that u(x, t) is an even function of x. Let w(x, t) = u(x, t) u(x, t). Then w(x, t) satises the wave equation, and we have the initial condition wave equation wtt = c2 wxx , w(x, 0) = 0, wt (x, 0) = 0. By uniqueness, w(x, t) 0.Then u(x, t) u(x, t) = 0 u(x, t) = u(x, t). Thus, u(x, t) is an even function of x.

#15) Use the energy method to prove the uniqueness of the diusion problem with Neumann boundary conditions: ut kuxx = f (x, t) for {0 < x < , 0 < t < } u(x, 0) = (x), ux (0, t) = g (t), ux ( , t) = h(t). Proof. Suppose u1 and u2 are solutions to the above equation with boundary conditions. Let w(x, t) = u1 (x, t) u2 (x, t). Then we have the following diusion problem with boundary conditions: wt kwxx = 0 for {0 < x < , 0 < t < } w(x, 0) = 0, w x (0, t) = 0, wx ( , t) = 0.

Then 0 = 0 w = (wt kwxx )w 1 0 = (w2 )t (kwx w)x + k (wx )2 2 Integrating both sides with respect to x, we have 0= 0= 1 2 (w2 )t dx (kwx w)|0 +
0 0

k (wx )2 dx

1 2 t

w2 dx +
0 0

k (wx )2 dx

Then t Then
0

w2 dx = 2
0 0

k (wx )2 dx 0

w2 (x, t)dx is decreasing with respect to t. Since t > 0, we have 0

0

w2 (x, t)dx
0

w2 (x, 0)dx = 0.

Then w2 (x, t) 0 w(x, t) 0 u1 (x, t) u2 (x, t) 0 u1 (x, t) u2 (x, t). Then, there is a unique solution.

HOMEWORK 5 SOLUTIONS

Section 3.1 : #1, 3, 4 #1) Solve the diusion equation on the half-line with the Dirichlet boundary condition: ut = kuxx on {0 < x < , 0 < t < } u(x, 0) = ex u(0, t) = 0 Solution: Using the method of odd extensions, consider the diusion equation on the whole line vt = kvxx on { < x < , 0 < t < } x if x > 0, e 0 if x = 0, v (x, 0) = odd (x) = x e if x < 0, The guarantees that v (x, t) is an odd function, and therefore v (0, t) = 0, and u(x, t) = v (x, t) when x > 0. The solution for v (x, t) was derived in 2.4. The result is

v (x, t) =

S (x y, t)odd (y )dy
0

=
0

S (x y, t)ey dy +
0

S (x y, t)(ey )dy S (x y, t)ey dy

0

=
0

S (x y, t)ey dy S (x y, t)ey dy +
0

= =
0

S (x + y, t)ey dy

[S (x y, t) S (x + y, t)] ey dy 1 4kt

e
0

(xy )2 4kt

(x+y )2 4kt

ey dy

If we want to write this in terms of E rf (x), we proceed as in problem 2.4.8 worked out in homework 4. The result is u(x, t) = 1 ktx e 1 E rf 2 2kt x 4kt 1 ekt+x 1 E rf 2 2kt + x 4kt

#3) Solve the diusion equation on the half-line with the Neumann boundary condition: wt = kwxx on {0 < x < , 0 < t < } w(x, 0) = (x) wx (0, t) = 0 Solution: Using the method of even extensions, consider the diusion equation on the whole line vt = kvxx on { < x < , 0 < t < } v (x, 0) = even (x) = (x) (x) if x 0, if x 0,

The guarantees that v (x, t) is an even function, and therefore vx (0, t) = 0, and w(x, t) = v (x, t) when x > 0. Once again, the solution for v (x, t) was derived in 2.4. The result is

v (x, t) =

S (x y, t)even (y )dy

=
0

S (x y, t)(y )dy
0

S (x y, t)(y )dy

After a change of variables in the second integral from y to y , we have

v (x, t) =
0

S (x y, t)(y )dy +
0

S (x + y, t)(y )dy

=
0

[S (x y, t) + S (x + y, t)] (y )dy 1 4kt

e
0

(xy )2 4kt

+ e

(x+y )2 4kt

(y )dy

#4) Solve the diusion equation on the half-line with the Robin boundary condition: ut = kuxx on {0 < x < , 0 < t < }, u(x, 0) = x for x > 0, ux (0, t) 2u(0, t) = 0 Solution: To solve the diusion Robin problem on the half-line, you take the same approach as solving the Dirichlet and Neumann problems. We consider a function dened on the whole line with an extension of the initial conditions that results in the boundary conditions being satised. We want to extend u(x, t) to a function v (x, t) that satises the diusion equation for all x, with v (x, 0) = 0 for x > 0. We also want vx (0, t) 2v (0, t) = 0. This

will be satised if we make sure that vx (x, t) 2v (x, t) is an odd function of x. Let f (x) = and v (x, t) = 1 4kt

x x + 1 e2x

for x > 0 for x < 0

(xy )2 4kt

f (y )dy

(a) You should notice that v (x, t) is the solution to the diusion equation on the whole line with initial condition v (x, 0) = f (x). If you do not see this, you will need to take the rst partial derivative with respect to t and the second partial derivative with respect to x. You will then need to nd out the value of v (x, 0). In the present form you cannot substitute t = 0. You will need to rewrite v (x, t) as: 1 v (x, t) =

ep f (x

4ktp)dp

by making a change of variables. vt = kvxx v (x, 0) = f (x) on { < x < , 0 < t < },

(b) Let w = vx 2v . Since w is a linear combination of v and vx , w also satises the diusion equation. wt = kwxx w(x, 0) = f (x) 2f (x) on { < x < , 0 < t < },

(c) Calculating f (x) 2f (x), we have f (x) 2f (x) = then f (x) 2f (x) = 1 + 2x x < 0 1 + 2x x > 0 1 2x x > 0 1 2x x < 0

So we have shown that f (x) 2f (x) = [f (x) 2f (x)]. Then f (x) 2f (x) is an odd function of x. (d) We know from 2.4.11 that any function that satises the diusion equation with an initial condition that is odd function of x, is also an odd function of x. (e) Since w(x, t) is an odd function of x, w(0, t) = 0 for t > 0. Then vx (0, t) 2v (0, t) = 0 Also, v (x, 0) = f (x) = x for x > 0. Then for x > 0, v (x, t) satises the diusion equation with given Robin boundary conditions. If the

solution is unique, it must be given by u(x, t) = 1 4kt

(xy )2 4kt

f (y )dy.

Section 3.2 : #1, 6 #1) Solve the Neumann problem for the wave equation on the half-line: utt = c2 uxx on x, t (0, ) u(x, 0) = (x) ut (x, 0) = (x) ux (0, t) = 0 Solution: Using the method of even extensions, consider the wave equation on the whole line vtt = c2 vxx on x, t (0, ) v (x, 0) = even (x) = vt (x, 0) = even (x) = (x) x 0 (x) x 0 (x) x 0 (x) x 0

The guarantees that v (x, t) is an even function, and therefore vx (0, t) = 0, and u(x, t) = v (x, t) when x > 0. The solution to the wave equation on the whole line is given by v (x, t) = 1 1 [even (x + ct) + even (x ct)] + 2 2c 1 1 [(x + ct) + (x ct)] + 2 2c
x+ct x+ct

even (y )dy
xct

When x ct > 0, even (x ct) = (x ct). Then we have

x+ct

u(x, t) =

(y )dy
xct

When 0 < x < ct, x ct < 0, so even (x ct) = (ct x). Then u(x, t) = = 1 1 [(x + ct) + (ct x)] + 2 2c 1 1 [(x + ct) + (ct x)] + 2 2c (y )dy +
0 x+ct

1 2c 1 2c

(y )dy
xct ctx

(y )dy +
0

(y )dy
0

#6) Solve the Robin problem for the wave equation on the half-line: utt = c2 uxx on x, t (0, ) u(x, 0) = 0 ut (x, 0) = V ut (0, t) + aux (0, t) = 0 With V, a, c > 0 and a > c.

Solution: Consider the function w(x, t) dened on the whole line by , for x > 0; ut (x, t) + aux (x, t) 0 , for x = 0; w(x, t) = ut (x, t) aux (x, t) , for x < 0. Since w(x, t) is a linear combination of derivatives of u(x, t), it also satises the wave equation. Next, we nd the initial conditions that w(x, t) satises. , for x > 0; ut (x, 0) + aux (x, 0) 0 , for x = 0; w(x, 0) = ut (x, 0) aux (x, 0) , for x < 0. , for x > 0; V + ax u(x, 0) 0 , for x = 0; = V ax u(x, 0) , for x < 0. V + ax (0) , for x > 0; 0 , for x = 0; = V ax (0) , for x < 0. , for x > 0; V 0 , for x = 0; w(x, 0) = V , for x < 0. We also have, wt (x, 0) = utt (x, t)|t=0 + auxt (x, t)|t=0 = c2 uxx (x, 0) + ax (ut (x, 0))
2 = c2 x (0) + ax (V )

wt (x, 0) = 0 Then we have the following wave equation on the whole line wtt = c2 wxx V, for x > 0 0, for x = 0 w(x, 0) = (x) = V, for x < 0 wt (x, 0) = 0 Then 1 [(x + ct) + (x ct)] 2 V, for x > ct; w(x, t) = 0, for 0 < x < ct. w(x, t) = Then for x > ct, we have ut (x, t) + aux (x, t) = V.

The homogeneous solution to this is uh (x, t) = f (at x), and a particular solution is up (x, t) = V t. Then u(x, t) = V t + f (at x) u(x, 0) = f (x) = 0 Then f (x) 0. Then u(x, t) = V t for x > ct. When 0 < x < ct, we have ut (x, t) + aux (x, t) = 0 Then u(x, t) = h(at x). We know this must satisfy the wave equation. utt (x, t) = c2 uxx (x, t) a2 h (at x) = c2 h (at x) (a2 c2 )h (at x) = 0 Then h (at x) = 0 h (at x) = K h(at x) = K (at x) Then u(x, t) = K (at x) for 0 < x < ct. To solve for K , we ensure that u continuous at x = ct. Then V t = u(ct, t) = K (at ct) V K= ac Then we have for a nal solution to the initial problem u(x, t) = V t,
atx ac V,

for x > ct; for 0 < x < ct.

Section 3.3 : #1, 3 #1) Solve the inhomogeneous diusion equation on the half-line with the Dirichelt boundary condition: ut kuxx = f (x, t) on {0 < x < , 0 < t < } u(x, 0) = (x) u(0, t) = 0

Solution: Using the method of reection, consider the inhomogeneous diusion equation on the whole line x > 0; f (x, t), 0, x = 0; vt kvxx = fodd (x, t) = f (x, t), x < 0. x > 0; (x), 0, x = 0; v (x, 0) = odd (x) = (x), x < 0. The guarantees that v (x, t) is an odd function, and therefore v (0, t) = 0, and u(x, t) = v (x, t) when x > 0. The solution for v (x, t) was derived in 3.3. The result is
t

v (x, t) =

S (x y, t)odd (y )dy +
0

S (x y, t s)fodd (y, s)dyds

=
0

S (x y, t)(y )dy +
t

S (x y, t)[(y )]dy
t 0

+
0 0

S (x y, t s)f (y, s)dyds +

0 t 0

S (x y, t s)[f (y, s)]dyds [S (x y, t s) S (x + y, t s)]f (y, s)dyds

0

=
0

[S (x y, t) S (x + y, t)](y )dy +

#3) Use the subtraction method to solve the inhomogeneous Neumann diusion problem on the half-line: wt = kwxx on {0 < x < , 0 < t < } w(x, 0) = (x) wx (0, t) = h(t) Solution: Let v (x, t) = w(x, t) xh(t). Then we have vt kvxx = xh (t) on {0 < x < , 0 < t < } v (x, 0) = (x) xh(0) vx (0, t) = 0 To solve this, we can use the method of even extensions. Let xh (t), x > 0; 0, x = 0; ut kuxx = feven (x) = xh (t), x < 0. (x) xh(0), x > 0; 0, x = 0; u(x, 0) = even (x) = (x) + xh(0), x < 0.

Then u(x, t) will be an even function of x, which guarantees u satises the boundary condition ux (0, t) = 0. Also, v (x, t) = u(x, t) for x > 0. The solution to the inhomogeneous diusion problem on the whole line is derived in 3.3, and it is given by equation (2). Then for x > 0,
t

v (x, t) =

S (x y, t)even (y )dy +
0

S (x y, t s)feven (y, s)dyds

=
0

[S (x y, t) + S (x + y, t)][(y ) yh(0)]dy
t

+
0 0

[S (x y, t s) + S (x + y, t s)][yh (s)]dyds

Then we have

w(x, t) =
0

[S (x y, t) + S (x + y, t)][(y ) yh(0)]dy
t

0 0

[S (x y, t s) + S (x + y, t s)]yh (s)dyds + xh(t)

HOMEWORK 6 SOLUTIONS
Section 3.2 : #8, 10, 11 #8) For the wave equation in the nite interval (0, ) with Dirichlet conditions, explain the solution formula within each diamond-shaped region. Solution: The equation and boundary conditions we are considering is equation below dened for x (0, ). utt = c2 uxx u(x, 0) = (x) ut (x, 0) = (x) u(0, t) = u( , t) = 0 We can extend this to the whole line by extending and in a way such that their extensions are odd functions of x around x = 0 and x = . Dene ext and ext by for 0 < x < ; (x) for < x < 0; ext (x) = (x) extended to be of period 2 . for 0 < x < ; (x) for < x < 0; ext (x) = (x) extended to be of period 2 . and consider vtt = c2 vxx v (x, 0) = ext (x) vt (x, 0) = ext (x) This is dened on the whole line and u(x, t) = v (x, t) for x (0, ). u(x, t) = 1 1 1 ext (x ct) + ext (x + ct) + 2 2 2c
x+ct xct

ext (y )dy

The values of ext and ext will depend on which diamond-shaped region (x, t) is in. To determine for yourself how these extended functions relate to the given functions and , you should work through several examples like the one done on pages 62 63. The results are CASE 1: (n, n) with n even: u(x, t) = 1 1 1 (x ct + n ) + (x + ct n ) + 2 2 2c
x+ctn

(y )dy
xct+n

CASE 2: (n + 1, n) with n even: 1 1 1 u(x, t) = (x + ct n ) + (x + ct n ) + 2 2 2c

1 x+ctn

(y )dy
x+ctn

CASE 3: (n, n + 1) with n even: u(x, t) = 1 1 1 (x ct + n ) (x ct + (n + 2) ) + 2 2 2c CASE 4: (n, n) with n odd: 1 1 1 u(x, t) = (x + ct (n 1) ) (x ct +(n +1) )+ 2 2 2c CASE 5: (n + 1, n) with n odd: u(x, t) = 1 1 1 (x ct + (n + 1) ) (x ct + (n + 1) ) + 2 2 2c CASE 6: (n, n + 1) with n odd: 1 1 1 u(x, t) = (x + ct (n 1) ) + (x + ct (n + 1) ) + 2 2 2c
x+ct(n+1) xct+(n+1) xct+(n+1) xct+(n+2)

(y )dy
xct+n

(y )dy
x+ct(n1)

(y )dy
xct+(n+1)

(y )dy
x+ct(n1)

#10) Solve the wave equation on the nite interval x (0, 2 ) with the given boundary conditions: utt = 9uxx in {0 < x < u(x, 0) = cos(x) ut (x, 0) = 0 ux (0, t) = 0 u( , t ) = 0 2 Solution: Consider the wave equation on the whole line vtt = 9vxx v (x, 0) = cos(x) vt (x, 0) = 0 Since cos(x) is even around x = 0, and odd around x = 2 , we will have that v (x, t) is an even function of x around x = 0, and v (x, t) is an odd function of x around x = 2 . This will guarantees that vx (0, t) = 0 and v ( 2 , t) = 0. (NOTE: you also need to be sure that vt (x, 0) is both even around x = 0, and odd around x = 2 ). Then u(x, t) = v (x, t) for 0 < x < 2. u(x, t) = 1 1 [cos(x + 3t) + cos(x 3t)] + 2 2c = cos(x) cos(3t)
x+3t

, 0 < t < } 2

0ds
x3t

#11) Solve the wave equation on the nite interval x (0, ) with the given boundary conditions: utt = c2 uxx in {0 < x < , 0 < t < } u(x, 0) = 0 ut (x, 0) = x u(0, t) = u( , t) = 0 Solution: Consider the wave equation on the whole line vtt = c2 vxx v (x, 0) = 0 vt (x, 0) = ext (x) = x for < x < with period 2 .

Since ext (x) is odd around x = 0 and x = , then v (x, t) will be as well. Then v (0, t) = v ( , t) = 0. Then u(x, t) = v (x, t) for x (0, ). u(x, t) = 1 2c
x+ct xct

ext (y )dy

For any point (x, t), in a diamond-shaped region , you can nd u(x, t) by using the results of problem 3.4.8 given above. For any initial point or point on the boundary, you can check that this solution will satisfy the given conditions.

Section 3.4 : #1, 6, 11, 12 #1) Solve utt = c2 uxx + xt u(x, 0) = 0 ut (x, 0) = 0 Solution: From equation (3), we have: u(x, t) = = =
0 t 0 t 0 x+c(ts)

1 2c 1 4c
t

ysdyds
xc(ts)

y 2 s|xc(ts) ds

x+c(ts)

xts xs2 ds 1 3 xt 6

#6) Derive the formula for the inhomogeneous wave equation by factoring the operator.

(a) We have utt c2 uxx = f (x, t) (t cx )(t + cx )u = f (x, t) Let v (x, t) = (t + cx )u(x, t) f (x, t) = (t cx )v (x, t) (b) We can solve for u in terms of v by using a change of coordinates. Let t = t + cx x = ct x Then t + cx 1 + c2 ct x x= 1 + c2 t= Then u x u t 1 + = v x t t t 1 + cc Integrating both sides with respect to t , we have ut = u(x, t) = = Let s= Then we have
t

1 1 + c2 1 1 + c2

v (x, r)dr
cx t

v
cx

cr x r + cx , 1 + c2 1 + c2

dr

r + cx 1 + c2

u(x, t) =
0

v (x ct + cs, s) ds

(c) Similarly, we can solve for v in terms of f by using a change of coordinates. The result is
t

v (x, t) =
0

f (x + ct cr, r)dr

(d) Substituting, part (c) into part (d), we have

t s

u(x, t) =
0 0

f (x ct + 2cs cr, r)drds =

1 2c

(e) Part (d) is a particular solution to the wave equation. To obtain the full solution, we need to add the homogeneous solution. Then u(x, t) = 1 1 [(x + ct) + (x ct)] + 2 2c
x+ct

(y )dy +
xct

1 2c

#11) Show by direct substitution that u(x, t) = h t x c for x < ct and u(x, t) = 0 for x ct solves the wave equation on the half-line with zero initial data and boundary condition u(0, t) = h(t). utt = c2 uxx u(x, 0) = 0 ut (x, 0) = 0 u(0, t) = h(t) Solution: When x = 0, we have x < ct, then u(0, t) = h(t). When t = 0, we have x > ct, then u(x, t) 0. Then u(x, 0) = 0 and ut (x, 0) = 0. Clearly, for x > ct, u(x, t) = 0 satises the wave equation. It only remains to check that u(x, t) satises the wave equation for x < ct. x utt = h (t ) c x 2 1 = c 2 h (t ) c c = c2 uxx #12) Solve the following using Greens Theorem vtt c2 vxx = f (x, t) in 0 < x < v (x, 0) = (x) vt (x, 0) = (x) v (0, t) = h(t) Solution: For x > ct, the domain of dependence does not intersect the boundary x = 0. Then we can use Greens Theorem exactly as done on pages 73 74. The result is the same as for the whole line u(x, t) = 1 1 [(x + ct) + (x ct)] + 2 2c
x+ct

(y )dy +
xct

1 2c

f dxdt

For x < ct, the domain of dependence reects o of the boundary x = 0. Then instead of being a triangle, like before, the domain of dependence is a quadrilateral. You should draw the domain of dependence. The four vertices are x0 (0, t0 ) (x0 , t0 ) c (ct0 x0 , 0) (x + ct0 , 0)

Then dene the four lines that go along the edges of the domain of dependence to be L0 = the line from (ct0 x0 , 0) to (x + ct0 , 0) L1 = the line from (x + ct0 , 0) to (x0 , t0 ) x0 L2 = the line from (x0 , t0 ) to (0, t0 ) c x0 L3 = the line from (0, t0 ) to (ct0 x0 , 0) c Then by Greens theorem, we have f dxdt =
L0 +L1 +L2 +L3

(c2 vx dt vt dx)

Compute each line integral separately, and then add them together.
x0 +ct0

(c2 vx dt vt dx) =
L0 ct0 x0

(x)dx

(c2 vx dt vt dx) = c
L1 L1

dv

= cv (x0 , t0 ) c(x0 + ct0 ) (c2 vx dt vt dx) = c

L2 L2

dv x0 + cv (x0 , t0 ) c

= ch t0

(c2 vx dt vt dx) = c
L3 L3

dv x0 c

= c(ct0 x0 ) ch t0 After adding these together, we have

x0 +ct0

f dxdt = 2cv (x0 , t0 )+c(ct0 x0 )c(x0 +ct0 )

ct0 x0

(x)dx2ch t0

x0 c

Then solving for v , we have v (x0 , t0 ) = 1 1 [(x0 +ct0 )(ct0 x0 )]+ 2 2c

x0 +ct0

(x)dx+
ct0 x0

1 2c

f dxdt+h t0

x0 c

Then the full solution for v is given by v (x, t) =

1 2 [(x 1 2 [(x

+ ct) (ct x)] + + ct) + (x ct)] +

1 2c 1 2c

x+ct ctx x+ct xct

(y )dy + (y )dy +

1 2c 1 2c

f dxdt + h t f dxdt

x c

if x < ct if x > ct

Section 4.1 : #4, 6 #4) Write the series expansion solution of a wave in a resistant medium that satises utt = c2 uxx rut u(x, 0) = (x) ut (x, 0) = (x) u(0, t) = u( , t) = 0 where r is a constant such that 0<r< 2c for 0 < x <

Solution: Write u(x, t) as separated u(x, t) = X (x)T (t) Substituting this into the dierential equation, we have X (x)T (t) = c2 X (x)T (t) rX (x)T (t) Then T (t) + rT (t) X (x) = = = 2 > 0 c2 T (t) X (x) Then we have two ODEs X (x) + 2 X (x) = 0 T (t) + rT (t) + c2 2 T (t) = 0 Solving the rst ODE, we have X (x) = C1 cos(x) + C2 sin(x) Since X (0) = X ( ) = 0, we have = Then Xn (x) = sin The second ODE is now T (t) + rT (t) + nc
2

which means n = nx

for n N.

T (t) = 0.

This has the characteristic equation nc s2 + rs + Solving this for s, we have r s= r2 4 2

=0

nc 2

Since 0 < r < T (t) = ARe e

2c

rt+t

, s is a complex number. Then ( nc )

2 rt+t

r 2 4 2

r 2 4 2

( nc )

+ B Im e
2

Tn (t) = An cos t un (x, t) = An cos t

r2 nc 4 r2 nc 4

+ Bn sin t
2

r2 nc 4 t r2 nc 4

e
2

rt 2

+ Bn sin
2

rt 2

sin

nx

u(x, t) =
n=1

An cos t

r2 nc 4

+ Bn sin t

r2 nc 4

rt 2

sin

nx

where

(x) =
n=1

An sin Bn
n=1

nx r2 nc 4

(x) =

r An 2

sin

nx

#6) Consider the following dierential equation on a nite interval with given boundary conditions. Separate variables and show that there are an innite number of solutions for the given initial condition. tut = uxx + 2u u(0, t) = u(, t) = 0 u(x, 0) = 0 Letting u(x, t) = X (x)T (t), we have tX (x)T (t) = X (x)T (t) + 2X (x)T (t) Then separating the variables, X (x) tT (t) 2= = T (t) X (x) Then we have two ODEs X (x) + X (x) = 0 tT (t) + ( 2)T (t) = 0 The rst ODE, yields X (x) = sin( x) Since X (0) = X ( ) = 0, we have Xn (x) = sin(nx) for n N

where n = n2 . The second ODE is separable, and we can rewrite it as T (t) 2 = T (t) t T (t) 2 dt = dt T (t) t ln(T (t)) = (2 ) ln(t) + C T (t) = Ct2 Tn (t) = Cn t2n Then u(x, t) =
n=1
2

Cn t2n sin(nx)

Since u(x, 0) = 0, we require that Cn = 0 for n 2. Then u(x, t) = C1 t sin(nx) However, we can choose any value we want for C1 , and u(x, t) will still satisfy the dierential equation, the boundary conditions, and the initial condition.

HOMEWORK 9

6.1.4. We look for a solution u = u(r). Thus u must satisfy 2 c1 urr + ur = 0 (r2 ur )r = 0 r2 ur = c1 u = + c2 . r r From the boundary conditions c1 + c2 = A a c1 u(b) = B + c2 = B. b Solving for c1 , c2 we obtain u(a) = A c1 = (B A)(1/a 1/b)1 , 6.1.11. Integrate u = f to obtain (1)
D

c2 = B + (B A)(1/a 1/b)1 /b.

udx =
D

f dx .

From the divergence theorem (2)

D

udx =
D

div(u)dx =
D

u ndS =
u n

Combining (1), (2) and the boundary condition desired equality.

u dS. n D = g on D, we obtain the

HOMEWORK 10

6.2.1. Following the hint, guess that u(x, y ) = Ax2 + By 2 + Cxy + Dx + Ey + F. Then, (1) u = 0 A = B. (2) ux (0, y ) = a CY + D = a. (3) ux (a, y ) = 0 2Aa + CY + D = 0. (4) uy (x, 0) = b Cx + E = b.

(5) uy (x, b) = 0 2Bb + Cx + E = 0. From (2) and (3) one immediately gets 2Aa a = 0 A = 1/2. From (4) and (5) one immediately gets 2Bb + b = 0 B = 1/2. This is consistent with (1). Since (2),(3), (4) and (5) must hold for all x, y , one must choose C = 0 which implies D = a, E = b. Thus u(x, y ) = 1/2x2 1/2y 2 ax + by + F. Note, this is a Neumann problem hence there is uniqueness up to additive constants. 6.3.1. (a) By the maximum principle u achieves its maximum on r = 2. Hence it suces to nd the maximum value of u = 3 sin 2 + 1. Clearly this max value is 4, as the sine function has 1 as its max value. (b) By the MVP, (a = 2) u(0) = 1 4
2

(3 sin 2 + 1)2d = 1.
0

6.4.6. We look for a separated solution u = R. Then (1) + = 0 (0) = ( ) = 0.

(2) r2 R + rR R = 0 R(0) f inite, R(1) = sin sin 2. From (1) we have (Dirichlet problem)
1

HOMEWORK 10

= n2 , n = sin n, n = 1, 2, . . . Substituting = n2 in (2) and looking for a solution of the form r , we obtain = n. Thus, Rn = Crn + D/rn . We choose D = 0, so that R is nite at zero. Then our separated solutions look like, un = An rn sin n. Summing them up we get,

u=
n=1

An rn sin n.

By imposing the boundary condition at 1 we have,

sin sin 2 =
n=1

An sin n.

This immediately implies A1 = , A2 = 1, An = 0 Thus, u = r sin r2 sin 2. 7.1.5. Let v = u w. Then (1) E [w] = E [u]
D

n = 1 , 2.

u vdx +
D

|v |2 dx +
D

hvdS.

Now apply G1 to u and v : v Since u = 0 and

D u n

u dS = u vdx + n D = h on D, we get (2)

D

v udx.
D

hdS =
D

u vdx.

Combining (1) and (2) we get, E [w] = E [u] +

D

|v |2 dx E [u].

HOMEWORK 11

7.2.2. Apply G2 with u = , v = 1/(4 |x|), over the domain D = B \ B . (B = unit ball centered at zero, B = ball of radius centered at zero.) Then v = 0 in D and we get 1 dx = 4 |x| (
D

v v )dS = n n ( v v )dS. n n

(
B

v v )dS + n n

The rst addend in the sum above is zero because 0 outside of B , and it is smooth in R3 . As 0 we then obtain (r = |x|, /n = /r) 1 dx = lim 0 4 |x|
r=

v v n n

dS = lim

r=

1 4r dS +

+ 1 4

1 4r n n

dS =

= lim

1 4r2

1 4r n

dS = lim
0

1 4 2

dS =

r=

r=

+ = (0), = lim 0 n the average of f on the boundary of the where for any function f , we denote by f ball B . 7.3.2. For notational simplicity I will take x0 = 0. Apply G2 to u, v with v = (4 |x|)1 , over the domain D = D \ B (B = ball of radius centered at zero.) Then v = 0, u = f in D , and we get 1 f dx = 4 |x| (u
D

v u v )dS = n n (u v u v )dS. n n

(u
D

v u v )dS + n n

As

0, using the same argument as in the previous problem, we then obtain (1)

v u 1 f dx = (h v )dS u(0), 4 | x | n n D D where we also used that u = h on D. Now lets write G(x, 0) = v (x) + H (x) with H harmonic in D. We apply G2 to u, H in the domain D to obtain
1

HOMEWORK 11

(2)

Hf dx =
D

(h

H u H )dS, n n

where we also used that u = h on D. Adding (1) and (2), and using that G 0 on D we obtain
D

Gf dx =
D

G dS u(0), n

which gives the desired claim. 7.4.9. We have 1 1 + , 4 |x x0 | 4 |x x 0| where x 0 is the reection of x0 across the plane ax + by + cz = 0. Thus, set n = (a, b, c), the normal vector to the plane, we have G(x, x0 ) = x 0 = x0 2n ax0 + by0 + cz0 . a2 + b2 + c2

Property 1 : the only singularity of G is at x0 , since x0 is not in D. Property 2: when x D, then |x x0 | = |x x 0 |. Hence G(x, x0 ) = 0. Property 3: G(x, x0 ) + D, since x0 is not in D.
1 4 |xx0 |

1 4 |xx0 |

which is harmonic everywhere in

7.4.22. Following the hint, consider the function v = u/y , which solves v = 0 in {y > 0}, v = h on {y = 0}.

Thus, according to formula (4) page 183 (modied for the 2D case), v (x, ) =
+

h( ) d ( x)2 + 2

(I am calling (x0 , y0 ) with (x, ) and (x, 0) with (, 0). This is just for elegance purposes.) Integrating in d between y and + we get (we can assume that u behaves like a constant at ): u(x, y ) = C = C lim But
R +

y R +

h( ) dd ( x)2 + 2 h( ) dd ( x)2 + 2

1 R+

h( ) dd ( x)2 + 2

HOMEWORK 11 + R

h( )
y

dd = ( x)2 + 2

1 + = h( )[log(( x)2 + R2 ) log(( x)2 + y 2 )]d = 2 1 + = h( )[log(( x)2 + R2 ) log R2 log(( x)2 + y 2 )]d 2 Here we used that in order for the Neumann problem to have a solution one must have (problem 6.1.11)
+

h( )d = 0.

Thus,
R +

h( ) dd ( x)2 + 2

= Hence,

1 2

h( )[log(( x)2 /R2 + 1) log(( x)2 + y 2 )]d

u(x, y ) = C lim 1 =C+ 2

1 R+
+

h( ) dd ( x)2 + 2

h( ) log(( x)2 + y 2 )d.

Partial Dierential Equations Math 442 C13/C14 Fall 2009 Homework 1 Solutions
1. Determine which of the following operators are linear: (a) Lu = uxx + uxy (b) Lu = uux (c) Lu = 4x2 uy 4y 2 uyy Solution: compute: (a) L(u + v ) = (u + v )xx + (u + v )xy = uxx + vxx + uxy + vxy = (uxx + uxy ) + (vxx + vxy ) = Lu + Lv, and L(u) = (u)xx + (u)xy = uxx + uxy = (uxx + uxy ) = Lu. (b) L(u) = (u)(u)x = u(ux ) = 2 ux and this is not equal to Lu if 2 = . (c) L(u + v ) = 4x2 (u + v )y 4y 2 (u + v )yy = 4x2 (uy + vy ) 4y 2 (uyy + vyy ) = 4x2 uy 4y 2 uyy + 4x2 vy 4y 2 vyy = Lu + Lv, and L(u) = 4x2 (u)y 4y 2 (u)yy = 4x2 uy 4y 2 ux = (4x2 uy 4y 2 uyy ) = Lu. 2. (Strauss, 1.1.4) Show that the dierence of two solutions to Lu = g is a solution to Lu = 0, when L is any linear operator. Solution: Assume that Lu = g and Lv = g . Dene w = u v , then Lw = L(u v ) = Lu Lv = g g = 0, where the second equality comes from L being linear. 3. Solve: 2ux + 3ut = 0, u(x, 0) = x2 . Solution: From the formula derived in class, we know the solution is of the form u(x, t) = f (3x 2t), for some undetermined function f . Using the initial condition, we obtain u(x, 0) = f (3x) = x2 , so that f (s) = s2 /9, or u(x, t) = (3x 2t)2 /9. We will see that the rst and third are linear and the second is not. For example, we

4. Consider the heat equation with initial condition given as ut = uxx , u(x, 0) = x + , where , are real numbers. Make an educated guess for the solution to this PDE and check that it is correct. Interpret this as a statement about the evolution of temperature in a 1D object. Solution: There are at least two ways to come up with a guess. If we think of the temperature of an object as being a linear function, notice then that at every point, we expect the heat coming into the point to be the same as the heat leaving, so while there is a heat ux at every point, the actual temperature stays xed. Thus we expect this temperature not to change, and we would guess u(x, t) = x + . Another approach is to notice that if we plug the initial condition into the right hand side of the equation, then we get zero, which means that ut is also zero, which means that u should not change in time. Then we would guess that u(x, t) = x + for all t. Either way, though, we can check that u(x, t) = x + is a solution to the system. Clearly it satises the initial condition u(x, 0) = x + , and moreover it satises the PDE (ut = 0 and uxx = 0). So it is a solution. (We will show later in class that it is the only solution.) 5. Determine the type of the following equations: (a) uxx + uxy + uyy + 3uy = 0, (b) 9uxx uy = 0. (c) Now, for the equation uxx + 3uxy + uyy = 0, determine which values of make the equation elliptic. Solution: (a) Using the notation of class (or the book), we have 11 = 1, 12 = 1 2 , 22 = 1, and thus 11 22 = 1 1 and 2 = , so the equation is elliptic. 12 4 (b) We have 11 = 9 and 12 = 22 = 0, and thus we have 0 9 = 02 , so that the equation is parabolic. In fact, writing this equation as uy = 9uxx and thinking of y as time, we have the heat equation exactly. (c) Here we have 11 = 1, 12 = equation is elliptic.
3 2.

So we compare 1 and (3/2)2 = 9/4. Thus if > 9/4, the

u 2u , xy by xy u = xy , and similarly for other 6. We dene the operator x by the equation x u = x independent variables. Moreover, when we concatenate operators, we take it to mean composition, i.e. LM u := L(M (u)).
2 (a) Show that x = xx .

(b) Show that (x + y )2 = 2 xx + 2xy + 2 yy , i.e. that they multiply like polynomials. (c) From this, prove that if , , are complex numbers, then there exist 1 , 2 complex such that xx + xy + yy = (x 1 y )(x 2 y ). Compute 1 , 2 in terms of , , . 2

Solution: We compute the eect of each operator on a function and see that it is the same. (a) We check:
2 x u = x x u = x

u 2u = xx u. = x x2

This is true for any function u and we are done. (b) Now we compute: (x + y )2 u = (x + y )(x + y )u = (x + y )(ux + uy ) = x (ux + uy ) + y (ux + uy ) = 2 uxx + uyx + uxy + 2 uyy = 2 uxx + 2uxy + 2 uyy = (2 xx + 2xy + 2 yy )u. (c) There are a couple of ways to do this, but the most straightforward might be to compute the right-hand side rst, and we have (x 1 y )(x 2 y )u = (x 1 y )(ux 2 uy ) = [x (ux 2 uy ) 1 y (ux 2 uy )] = [uxx (1 + 2 )uxy + 1 2 uyy ] . So this will work, if we have (1 + 2 ) = , 1 2 = Solving the second equation for 1 gives 1 = /2 ; plugging this into the rst gives + 2 = , 2 or 2 2 + 2 + = 0. (Notice this similarity of this equation and the original left-hand side of the equation.) We then have two solutions for this quadratic polynomial, namely 2 = 2 4 , 2

and its not hard to see that if we pick 2 to be either of these roots, this will make 1 the other one.

Partial Dierential Equations Math 442 C13/C14 Fall 2009 Homework 2 due September 18
1. (Strauss 2.1.1.) Solve utt = c2 uxx , u(x, 0) = ex , ut (x, 0) = sin x. Solution: We use dAlemberts formula with (x) = ex , Then we have
x+ct

(x) = sin x.

sin s ds = cos(x + ct) cos(x ct),

xct

and so we have u(x, t) =

1 1 x+ct (e + exct ) + (cos(x + ct) cos(x ct)). 2 2c

2. (Strauss 2.1.7.) We dene an odd function to be any function f such that f (x) = f (x) for all x. Prove that if that the initial conditions , are odd functions, then so is the solution u(x, t) for any xed time t. Solution: There are two completely dierent solutions: Again use dAlembert. We know that the solution is u(x, t) = Now, we compute u(x, t) = 1 1 ((x + ct) + (x ct)) + 2 2c
x+ct

1 1 ((x + ct) + (x ct)) + 2 2c

x+ct

sin s ds.
xct

(s) ds
xct

Using the fact that is odd, we have (x ct) = ((x + ct)) = (x + ct), (x + ct) = ((x ct)) = (x ct). Now consider the integral, and make a change of variables r = s, giving
x+ct xct x+ct

(s) ds =
xct x+ct

(r) dr =
xct

(r) dr.

(Note in the last equality there are actually three minus signs, since we use the fact that is odd, and we ip the domain of integration.) Putting all of this together gives u(x, t) = u(x, t) and thus u is odd. We showed that reversing time gives a solution to the wave equation, and so does reversing space: if we choose v (x, t) = u(x, t), then we see that vtt (x, t) = utt (x, t), and thus if u satises utt = c2 uxx , u(x, 0) = (x), 1 ut (x, 0) = (x), vxx (x, t) = uxx(x, t),

then v satises vtt = c2 vxx , v (x, 0) = (x), vt (x, 0) = (x). This is true for any solution to the wave equation. If we now further assume that , are odd, then these two PDE have the same initial data, and therefore by uniqueness, v (x, t) = u(x, t) for all x, t, and thus u(x, t) = u(x, t), and u is odd. 3. We have dened a well-posed problem in class (also see book) typically for PDE, but we can consider if an ODE satises these three properties as well. Here you are given a sequence of ODEs and initial conditions; determine which of these problems are well-posed, and which are not1 : dy = 2y, y (0) = 2, dt dy (b) = ln x, y (0) = 0. dx (a) Solution: (a) This ODE satises the hypotheses of the E-U theorem for ODEs (in fact, the vector eld is C ) and thus it is well-posed. (b) We cannot apply the theorem directly, but we can solve this ODE exactly. In fact, the general solution can be written as y (x) = x log x x + C for some constant C . We see that there is exactly one solution with y (0) = 0 (in fact, choose C = 0). To show stability, we need to show that if we choose two dierent initial conditions, we can make the solutions close by choose the initial conditions close. So consider the two solutions y1 (x) = x log x x, y2 (x) = x log x x + C.

Clearly, y2 (0) = C , and by choosing |y2 (0)| < , we can make |y1 (x) y2 (x)| < for any x. 4. (Strauss 2.2.2.) Let us consider a solution to the wave equation utt = uxx (we have assumed that 1 2 c2 = 1). Dene the energy density e(x, t) = 2 (u2 t + ux ) and the momentum density p(x, t) = ut ux . Show that p p e e = and = , t x t x (b) e and p both satisfy the wave equation themselves (although with dierent initial conditions). (a) Solution: (a) We have e p = ut utt + ux uxt , = utx ux + ut uxx . t x shows these are equal. (similar for the other)

Using utt = uxx (b) We have

ex = ut utx + ux uxx ,
2 exx = u2 tx + ut utxx + uxx + ux uxxx ,

et = ut utt + ux uxt ,
2 ett = u2 tt + ut uttt + utx + ux uxtt
1 Recall

the ExistenceUniqueness Theorem which you saw in ODEs

Then
2 exx ett = ut utxx ut uttt + u2 xx utt + ux uxxx ux uttx 2 = ut (uxx utt )t + u2 xx utt + ux (uxx utt )x = 0 + 0 + 0 = 0.

(similar for p) 5. (Strauss 2.2.3.) Show the following invariance properties for solutions of the wave equation. Assume that u(x, t) satises the wave equation, then show that each of the transformed solutions also satisfy the wave equation: (a) translation: u(x , t) for any , (b) derivative: ux (x, t), (c) dilation: u(ax, at) for any a Solution: (a) Dene v (x, t) = u(x , t). We see, using the chain rule, that u d u v (x, t) = (x , t) (x ) = (x , t). x x dx x Similarly, 2u 2v ( x, t ) = (x , t). x2 x2 We also work out that 2u 2v ( x, t ) = (x , t). t2 t2

Therefore vtt (x, t) c2 vxx (x, t) = utt (x , t) c2 uxx (x , t) = 0. (b) Dene v (x, t) = ux (x, t). Then vtt = uxtt , so vtt c2 vxx = uxtt c2 uxxx = uttx c2 uxxx = (utt c2 uxx )x = (c) Dene v (x, t) = u(ax, at). Then v (x, t) = x 2v (x, t) = x2 v (x, t) = t 2v (x, t) = t2 Thus we have vtt (x, t) c2 vxx (x, t) = a2 utt (ax, at) a2 c2 uxx (ax, at) = a2 (utt (ax, at) c2 uxx (ax, at)) = a2 0 = 0. u (ax, at) a, x u 2u 2u (a (ax, at)) = a 2 (ax, at) a = a2 2 (ax, at), x x x x u (ax, at) a, t u 2u 2u (a (ax, at)) = a 2 (ax, at) a = a2 2 (ax, at). t t t t vxx = uxxx, 0 = 0. x

6. (Strauss 2.3.3.) Consider a solution to the diusion equation ut = uxx for x [0, L] and t > 0. Dene M (T ) = maximum of u(x, t) on the rectangle [0, L] [0, T ], m(T ) = minimum of u(x, t) on the rectangle [0, L] [0, T ]. Does M (T ) increase or decrease as a function of T ? Does m(T ) increase or decrease as a function of T ? Explain why. Solution: It turns out the answer to this is somewhat complicated as it could depend on the boundary and initial conditions. Let us rst x the boundary conditions as u(0, t) = 0, u(L, t) = 0,

for all t and assume that the initial condition (x) is positive for some x [0, L] so that M (0) > 0. The claim here then is that M (T ), m(T ) are functions which are constant in T . First, notice that by denition, if T > T , then M (T ) M (T ) (since were taking the maximum over a larger set). We now want to prove that M (T ) M (T ), and then we are done. So we prove by contradiction: assume that M (T ) > M (T ). If this is so, then clearly the maximum inside the rectangle [0, L] [0, T ] must occur for t (T, T ]. Since M (T ) > M (T ) M (0) > 0, this means that this maximum may not occur on the left- or right-hand edges, but must occur either in the interior of the rectangle, or on the top edge. But this directly contradicts the Maximum Principle. Since assuming M (T ) > M (T ) leads to a contradiction, it must be true that M (T ) M (T ). The argument for m(T ) is similar: reverse every inequality above and use the Minimum Principle. Of course, with dierent boundary conditions things could be more complicated. If we now assume that u(0, t) = f (t), and this function is increasing (rapidly enough), then its possible that M (T ) increases, since M (T ) maxt[0,T ] f (t) at least. In this case, the nal statement would be that M (T ) can increase, but no faster than f . Other permutations are left to the reader...

Partial Dierential Equations Math 442 C13/C14 Fall 2009 Homework 3 Solutions
1. Here we will prove that solutions to the heat equation satisfy (some of) the invariance principles mentioned in class, or in the book in 2.4. That is, if u(x, t) is a solution to ut = kuxx for x R, t > 0, then so are (a) u(x y, t) for any xed y , (b) ux , ut , u(x y, t)g (y ) dy where g has nite support, (d) v (x, t) = u( ax, at) for any a > 0. Solution: (a) Let v (x, t) = u(x y, t). Then v (x, t) = t v (x, t) = x 2v (x, t) = x2 Then u u (x y, t) 1 = (x y, t), t t u u (x y, t) 1 = (x y, t), x x v u 2u (x y, t). (x, t) = (x y, t) = x x x x x2 (c) v (x, t) =

v u 2v 2u (x, t) k 2 (x, t) = (x y, t) k 2 (x y, t) = 0, t x t x since u solves the heat equation. (b) We compute for ux , the other is similar. Denoting v = ux gives vt = (ux )t = uxt , vxx = (ux )xx = uxxx . Then vt kvxx = uxt uxxx = utx uxxx = (ut uxx )x = 0x = 0. (c) Since g has compact support, we can exchange derivatives and integration (see e.g. Theorem A.3.2 from Strauss), and thus we have v = t t and 2v 2 = 2 x x2

u(x y, t)g (y ) dy =

u (x y, t)g (y ) dy t 2u (x y, t)g (y ) dy. x2

u(x y, t)g (y ) dy =

But then vt kvxx =

u (x y, t)g (y ) dy t

2u (x y, t)g (y ) dy x2

u 2u (x y, t) 2 (x y, t) g (y ) dy = t x 1

0 dy = 0.

(d) We have u v (x, t) = a ( ax, at), t t u v (x, t) = a ( ax, at), x x 2v 2u (x, t) = a 2 ( ax, at). x2 x So then vt kvxx = aut ( ax, at) auxx( ax, at) = a(ut ( ax, at) uxx ( ax, at)) = a 0 = 0.

2. (Strauss 2.4.1.) Solve the heat equation with initial condition (x) = 1, |x| < L, 0, |x| L.

(You can use the formula for the solution as derived in class, but there is a simpler way to build this solution using the invariance principles above.) Solution: We will solve two ways, the rst using the formula. We have

u(x, t) =

S (x y, t)(y ) dy,

where S (x, t) = We can then write

L L

1 x2 /4kt e . 4kt

u(x, t) =
L

S (x y, t)(y ) dy +
L L

S (x y, t)(y ) dy +
L

S (x y, t)(y ) dy

=
L

0 dy +
L

S (x y, t) dy +
L

0 dy

=
L

S (x y, t) dy,

L 2 1 e(xy) /4kt dy. 4kt L Changing variables with s = (x y )/ 4kt, ds = 1/ 4kt dy , gives

so we need to evaluate

1 u(x, t) = 4kt 1 = = 1 2

x L 4kt x+ L 4kt

2 es ( 4kt) ds
2

x+ L 4kt x L 4kt

es ds erf xL 4kt .

erf

x+L 4kt 2

A completely dierent method is to use the transformations in Question #1, particularly (a). We know that if the initial condition is the Heaviside function H (x), then we get the solution 1 1 Q(x, t) = + 2
x/ 4kt 0

ep dp =

1 2

1 + erf

x 4kt

We can see that our initial condition (x) can be written (x) = H (x + L) H (x L). (One can either draw this, or check it algebraically: if x > L, then H (x + L) = H (x L) = 1, if x < L, then H (x + L) = H (x L) = 0, and if L < x < L, then H (x L) = 0 but H (x + L) = 1.) From 1(a) above, we know that Q(x + L, t) and Q(x L, t) are both solutions to the heat equation, and clearly they have initial conditions H (x + L) and H (x L), respectively. By linearity, we know that 1 1 x+L xL 1 1 erf u(x, t) = Q(x + L, t) Q(x L, t) = + erf 2 2 2 2 4kt 4kt is also a solution, and it clearly satises u(x, 0) = (x). 3. (Strauss 2.4.8.) Show that the tails of
2 1 ex /4kt S (x, t) = 2 kt

are uniformly small for small times, i.e. that for any > 0,
t0 |x|>

lim max S (x, t) = 0.

Interpret this in terms of speed of propagation of information for solutions of the heat equation. Solution: We rst compute the inner term, namely
|x|>

max S (x, t).

First, note that for x > 0 and any xed t > 0, S (x, t) is monotone decreasing, because
2 S 2x (x, t) = ex /4kt < 0. x (4kt)3/2

Therefore,

2 1 e /4kt . max S (x, t) = S (, t) = x> 4kt

By evenness of S , we also have that maxx< S (x, t) = maxx> S (x, t), so

2 1 max S (x, t) = e /4kt . 4kt

|x|>

So it remains to compute e /4kt lim . t0 4kt This is an indeterminate form, since when we plug in t = 0 we obtain 0/0. The rst guess might be to try lHopitals rule, but this will actually not work out, because every time we dierentiate the numerator, we will get a higher power of t in the denominator. 3
2

To simplify the expression, rewrite this limit as eC1 /t t0 C2 t lim and make the change of variables s = 1/t, which then gives C 1 s s se = lim . lim C s C2 e 1 s s C2 This is still an indeterminate form of /, but now using lHopitals Rule gives s1/2 /2 0 = = 0. s C1 C2 eC1 s lim

4. (Strauss 2.4.9.) We will write down an exact solution to the heat equation ut = kuxx , u(x, 0) = x2 , but not using the formula derived in class. The idea is as follows. (a) Show that uxxx solves the heat equation with initial condition zero, (b) Use uniqueness to show uxxx (x, t) 0, (c) From this we can deduce that u(x, t) = A(t)x2 + B (t)x + C (t) for some functions A, B, C (Why?), (d) Solve for A, B, C . Solution: (a) This part is similar to problem #1. If we write v = uxxx , then vt = uxxxt = utxxx , vxx = uxxxxx, and thus vt kvxx = utxxx kuxxxxx = (ut kuxx )xxx = 0xxx = 0. Moreover, notice that v (x, 0) = uxxx(x, 0) = (x2 )xxx = 0. (b) We know solutions to the heat equation are unique. Moreover, we know that v solves the heat equation with v (x, 0) = 0. However, it is easy to see that if we dene w(x, t) 0 for all x, t, then w satises the heat equation and w(x, 0) = 0. Therefore v w and v (x, t) 0. (c) We know that uxxx (x, t) = 0, uxx (x, t) = A(t), ux (x, t) = A(t)x + B (t), 1 u(x, t) = A(t)x2 + B (t)x + C (t), 2 where A(t), B (t), C (t) are arbitrary functions of t. Now redene A to get rid of the 1/2 since its arbitrary anyway. 4

(d) We know that u satises the heat equation, so we have ut = A (t)x2 + B (t)x + C (t), uxx = 2A(t), and if these are equal as functions, this gives C (t) = 2A(t), B (t) = 0, A (t) = 0.

Solving the last two are easy (A(t) = A0 ,B (t) = B0 ) and then the rst becomes C (t) = 2A0 t + C0 . Putting this together gives u(x, t) = A0 x2 + B0 x + (2A0 t + C0 ). Plugging in the initial condition gives u(x, 0) = A0 x2 + B0 x + C0 , which means that A0 = 1, B0 = C0 = 0, so the solution is u(x, t) = x2 + 2t. We could, alternately, plug in the initial conditions as soon as we have the equations, namely say that A (t) = 0, B (t) = 0, A(0) = 1, B (0) = 0, C (0) = 0,

C (t) = 2A(t), and directly solve to get A(t) = 1 and C (t) = 2t.

5. Generalize the previous problem to a general initial condition which is a polynomial of x. (You dont need to compute anything exactly here, just describe the algorithm which would allow you to obtain a solution.) Solution: The general idea is as follows. Lets say that u(x, 0) = p(x), where p(x) is a polynomial of degree n:
n

p(x) =
k=0

k xk .

Then notice that if we take n + 1 derivatives of p we get zero. Therefore, if u satises the heat equation n+1 u solves the heat equation with initial condition zero. Therefore with initial condition p(x), then xn+1 n+1 u 0, and by the same argument we know we have xn+1
n

u(x, t) =
k=0

Ak (t)xk .

Plugging this into the heat equation gives

n

ut (x, t) =
k=0 n

k A k (t)x , n2

uxx (x, t) =
k=0

k (k 1)Ak (t)xk2 =
k=0

(k + 2)(k + 1)Ak+2 (t)xk .

Setting these equal gives

A n (t) = An1 (t) = 0,

A k (t) = (k + 2)(k + 1)Ak+2 (t),

and plugging in initial conditions gives Ak (0) = k for all k . The rst two equations can be solved easily: An (t) = n , An1 (t) = n1 , and then the other equations can be solved recursively, e.g. A n2 (t) = n(n 1)An (t) = n(n 1)n , so An2 (t) = n(n 1)n t + n2 , etc.

Partial Dierential Equations Math 442 C13/C14 Fall 2009 Homework 4 due October 9
1. Consider the boundary value problem A + A = 0, A (0) + aA(0) = 0, A(L) = 0.

(a) Show that if a < 0, then there is no negative eigenvalue. (b) Under which conditions is there a zero eigenvalue? (c) Show there are innitely many positive eigenvalues for any value of a. Bonus: We showed in (a) that if a < 0 then there is no negative eigenvalue. If turns out that for some positive a, this problem has a negative eigenvalue (and for some others it does not). Write down a condition on a which determines whether such an eigenvalue exists. Solution. (a) Lets say we have a negative eigenvalue < 0. Then if we say A(x) = erx , we have r 2 + = 0, r = . Since < 0, this means the roots are real, let us write them as r = b where b > 0. Then we have A(x) = C1 ebx + C2 ebx . Plugging in the initial conditions gives C1 b C2 b + a(C1 + C2 ) = 0, C1 ebL + C2 ebL = 0. The second equation can be written as C2 = C1 e2bL , and plugging this into the rst gives C1 (a + b) + C1 e2bL (b a) = C1 (b(1 + e2bL ) + a(1 e2bL )). This must be zero, but C1 cannot be zero (since then so would C2 be), so we need to solve b(1 + e2bL ) + a(1 e2bL ) = 0. (1)

However, this equation has no solution! Notice that since bL > 0, we know e2bL > 1, and thus 1 e2bL < 0. Since a is also less than zero, we know that a(1 e2bL ) > 0. But since b(1 + e2bL ) > 0 as well, it is not possible to solve (1). (b) If we have a zero eigenvalue, this means we have a solution to A (x) = 0 with those boundary conditions. However, this means that A(x) = x + , and plugging this into the boundary conditions gives + a = 0, L + = 0. 1

We want to nd a nontrivial solution to this system, and if we write it as a matrix equation: 1 a L 1 = 0 0 ,

and this has a nontrivial solution i the matrix is singular, i.e. if its determinant is zero. So the condition is that 1 a det = 1 aL = 0, L 1 or aL = 1. (c) If > 0, then the solutions to the ODE look like C1 cos(x) + C2 sin(x), where 2 = . Plugging in the boundary conditions gives C2 + aC1 = 0, C1 cos(L) + C2 sin(L) = 0. Solving the rst equation gives C2 = aC1 / , and plugging this into the second and doing some algebra gives tan(L) = . a The question is, how many roots does this equation have? We cannot answer this analytically, but we can see from the graph that it has to have innitely many; see an example picture in Figure 1.
15

10

5 5

10

15

Figure 1: Graphs of tan(x) and x To prove this analytically, choose any nonnegative integer k and consider the interval [(k + 1/2)/L, (k + 3/2)/L]. We know from properties of tan that
/2) + t (k+1 L

lim

tan(x) = ,

/2) t (k+3 L

lim

tan(x) = ,

so for x slightly larger than (k +1/2)/L, tan(x) < x/a and for x slightly smaller than (k +3/2)/L, tan(x) > x/a. By the Intermediate Value Theorem, the functions must be share at least one point in the strip. Since the curves intersect in each of these strips, there are innitely many intersections. 2

Bonus: Repeating some of part (a), we know we need to solve (1) for b, given a positive a, so we have tanh(2bL) = a . b

These curves will always intersect as long as a > 0. Notice that tanh(0) = 0 and limb tanh(2bL) = 1, whereas limb0+ a/b = and limb a/b = 0, so the curves must cross. So the condition is that for any a > 0, there is a negative eigenvalue. 2. (Strauss 4.3.2.) Consider the eigenvalue problem with Robin boundary conditions A + A = 0, A (0) 0 A(0) = 0, A (L) + L A(L) = 0.

(a) Show that zero is an eigenvalue if and only if 0 + L = 0 L L. (b) Compute the eigenfunction corresponding to this eigenvalue. Solution. (a) If we have a zero eigenvalue, then we have A = 0 or A(x) = x + . Plugging in the boundary conditions gives 0 = 0, + L (L + ) = 0. We want to nd a nontrivial solution for , in this equation, or, as above, we need the matrix 1 0 1 + L L L to have determinant zero, or L + 0 (1 + L L) = 0. (b) If this determinant is zero, we know that the two equations we have are redundant, so we can solve either. The simpler to solve is the rst, which gives = 0 , and of course we will have one free choice for . So one eigenfunction we can choose is A(x) = 0 x + 1, and we can of course choose any scalar multiple of this. 3. Solve the equation ut = kuxx , u(x, 0) = u(0, t) = 0. x [0, ), t > 0,

1, x (0, 1), 0, x > 1,

Solution. We use the formula as derived in class (equation (6) in 3.1): 1 u(x, t) = 4kt

e (x y )
0

/4kt

e ( x +y )

/4kt

(y ) dy,

where (y ) = 1 for y < 1 and 0 otherwise. So we can also write 1 u(x, t) = 4kt Let us consider the rst term alone,
1 2 1 e(xy) /4kt dy. 4kt 0 Using the change of variables s = (x y )/ 4kt, we obtain 1

e (x y )
0

/4kt

e ( x +y )

/4kt

dy.

1 or 1 2

x 4kt x 1 4kt

es ds,

erf

x 4kt 1
1

erf

x1 4kt
2

Now consider the second term, e(x+y) /4kt dy. 4kt 0 Using the change of variables s = (x + y )/ 4kt, we obtain 1 or 1 2
x+1 4kt x 4kt

es ds,

erf

x+1 4kt

erf

x 4kt

Adding these together gives 1 2 erf x+1 4kt 2 erf x 4kt + erf x1 4kt .

4. Consider the Schr odinger equation with Neumann boundary conditions: iut = uxx , u u (0, t) = (L, t) = 0. x x

Write out the general series solution for this equation as we have done for the heat and wave equations, i.e. separate variables, get ODEs in x and t, solve these problems, and take the linear combination. Solution. We separate variables as usual, writing u(x, t) = A(x)B (t). Working this out gives us A + A = 0, A (0) = A (L) = 0, B + iB = 0.

for the A equation. Weve solved this in class, and we know that the eigenvalues and eigenfunctions are n n 2 , An (x) = cos x , n = 0, 1, 2, . . . n = L L 4

Solving the B equation gives Bn (t) = Bn (0)ein t , which oscillates. Notice of course that we can allow Bn (0) to be complex with no extra diculties. Thus we have un (x, t) = Cn ein t cos and we form a general solution by linear combinations:

n x , L

u(x, t) =
n=0

Cn ein t cos

n x . L

Partial Dierential Equations Math 442 C13/C14 Fall 2009 Homework 5 Solutions
1. (Strauss 5.2.2.) Show that cos(x) + cos(x) is periodic if is a rational number and compute its period. What happens if is not rational? Solution: Let us rst notice that if f, g are both periodic with period p, then so is their sum: (f + g )(x + p) = f (x + p) + g (x + p) = f (x) + g (x) = (f + g )(x). So we simply have to show that cos(x), cos(x) share a period. Note that the periods of cos(x) are 2n with n Z, and the periods of cos(x) are 2m/, where m Z. The question is then, do these two sets of numbers share an element, i.e. is there an n and an m so that 2n = If = p/q , then we have 2n = 2m ?

mq 2mq , or n = . p p

This has many solutions, choose, for example, m = p, n = q . If is irrational, then this does not work; in fact, solving the rst equation gives = which is only possible if is rational. 2. Dene f (x) = x3 on the interval [0, 1]. Compute its Fourier sine series and its Fourier cosine series. Solution: The Fourier sine series coecients are given by Bn = 2 L
L

m n

f (x) sin(nx/L) dx,

0

and in this case L = 1, so we need to compute

1

Bn = 2
0

x3 sin(nx) dx.

Integrating by parts several times gives Bn = and then the F.S.S. is

n=1

(1)n 2(6 n2 2 ) , n3 3

Bn sin(nx).
1

The Fourier cosine series coecients are given by An = 2

0

x3 cos(nx) dx.

Integrating by parts several times gives An = 6 + (1)n (3n2 2 6) , n4 4 1

and then the F.C.S. is

A0 + An cos(nx). 2 n=1

3. Consider the function f (x) = x on the interval [, ]. Compute the full Fourier series for f (x). Use Parsevals Identity to compute 1 . 2 n n=1 Solution: Since f is odd, we know all of the cosine terms will be zero, so we compute Bn = Parsevals Identity tells us that
2 Bn sin(nx), sin(nx) = f, f , n=1

x sin(nx) dx =

2(1)n+1 . n

or

4 = n2 n=1 Dividing gives

x2 dx =

2 3 . 3

2 1 = . n2 6 n=1

4. Solve the heat equation given by ut = kuxx, x [0, L], t > 0,

u(x, 0) = x, u(0, t) = u(L, t) = 0.

Solution: We make the Ansatz u(x, t) = A(x)B (t), and plugging in this gives us A(x)B (t) = kA (x)B (t), or B (t) A (x) = = . A(x) kB (t) A (x) + A(x) = 0, and we know the solutions of this are An (x) = sin(nx/L), n = 2 n2 2 . L2 A(0) = A(L) = 0,

Thus we rst solve

Then we have Bn = Cn ekn t , and our general solution is u(x, t) =

n

Cn ekn t sin(nx/L).

Plugging in t = 0 gives x=
n

Cn sin(nx/L),

and thus Cn are the Fourier sine series coecients of x, or Cn = Therefore we have u(x, t) =
n

2 L

x sin(nx/L) dx =
0

2(1)n+1 L . n

2(1)n+1 L kn2 2 t/L2 e sin(nx/L). n

5. (Strauss 5.3.8.) Let f and g satisfy the same Robin boundary condition at x = 0 and the same Robin boundary condition at x = L (i.e., we assume that f (0) + f (0) = g (0) + g (0) = f (L) + f (L) = g (L) + g (L) = 0.) Prove then that (f (x)g (x) f (x)g (x))|x=0 = 0. Deduce from this that eigenfunctions of a Robin BVP are orthogonal. Solution: We have f (0) = f (0), g (0) = g (0), f (L) = f (L), g (L) = g (L).
x =L

Plugging this into the left hand side gives f (L)g (L) f (L)(g (L)) (f (0)g (0) f (0)(g (0))) = 0.

6. Prove that if f has period p, then

p+a

f (y ) dy
a

is independent of a. Solution: One way to do this is to write

p+a p a p+a

=
a 0

+
p

(1)

and we have that

p+a a a

f (y ) dy =
p 0

f (y p) dy =
0

f (y ) dy,

(the rst is from the u-substitution z = y + p and the second is from periodicity) and thus the second and third terms in (1) cancel.

7. Consider the innite list of functions {1, cos(x), cos(2x), . . . , cos(nx), . . . , sin(x), sin(2x), . . . , sin(nx), . . . }. Show that this is an orthogonal set of functions on the set [, ], i.e. if we dene the inner product

f, g :=

f (x)g (x) dx,

then if we choose any two dierent functions from that list, then their inner product is zero. Solution: We rst notice that

sin(nx) cos(mx) dx = 0

since sin is odd and cos is even and thus their product is odd, and the integral of an odd function on a symmetric interval is always zero. Now, consider

sin(nx) sin(mx) dx =

cos((n m)x) cos((n + m)x) dx sin((n m)x) sin((n + m)x) + nm n+m

x =

,
x =

as long as n = m, and this is clearly zero by periodicity of sin. We also have

cos(nx) cos(mx) dx =

cos((n m)x) + cos((n + m)x) dx sin((n m)x) sin((n + m)x) nm n+m

x =

,
x =

as long as n = m, and this is again zero by periodicity of sin. Finally, to compute the inner products of the functions with themselves, choose n > 0, and notice that sin2 (nx) + cos2 (nx) = 1, and thus

sin2 (nx) + cos2 (nx) dx = 2.

Moreover, since sine and cosine are the same function under a phase shift, we know from the previous problem that

sin2 (nx) =

cos2 (nx)

and therefore they each equal . For n = 0 we have

cos2 (0x) dx =

dx = 2.

8. Let {fn (x)} be any sequence of functions which converge to f (x) uniformly on [a, b]. Prove then that fn (x) converge to f in the L2 sense as well. Show a counterexample to demonstrate that the converse is false, i.e. that we can have L2 converge but not uniform. (The term used for this is that uniform convergence is stronger than L2 ). 4

Solution: To prove the rst statement, notice that

b a

|fn (x)| dx
a

b x[a,b]

max f (x)

dx = max f (x) (b a).

x[a,b]

Then if
n x[a,b]

lim max f (x) = 0,

2

then clearly
b n

lim

|fn (x)|2 dx (b a)

n x[a,b]

lim max f (x)

=0

as well. Also, notice that it is nonnegative by denition, and therefore the limit is zero.

Partial Dierential Equations Math 442 C13/C14 Fall 2009 Homework 6 Solutions
1. Let D = {(x, y ) : x2 + y 2 < 4}, and solve u = 0, in D, u = 3 2 cos ,

r = 2.

Solution: Using the formula generated in class, we have that u(r, ) = A0 + rn (An cos(n) + Bn sin(n)), 2 n=1

for some unknown An , Bn . Plugging in the boundary condition r = 2 gives u(r, 2) = A0 + 2n (An cos(n) + Bn sin(n)) = 3 2 cos , 2 n=1

from which we can see that A0 = 6, A1 = 1, and all the rest of the coecients are zero, and thus the solution is u(r, ) = 3 + 2r cos().

2. Let D be the square [0, 1]2 . Solve u = 0 subject to the boundary conditions u(0, y ) = 0, ux (1, y ) = 0, u(x, 0) = x2 2x, uy (x, 1) = 0.

Solution: First make the Ansatz u(x, y ) = A(x)B (y ), separating as usual gives A (x) + A(x) = 0, B (y ) B (y ) = 0, Solving the rst system gives An (x) = sin(n x),
2 n = (n + 1/2), n = 1, 2, 3, . . . , n = n .

A(0) = A (1) = 0, B (1) = 0.

From this we obt ? 33T E < 8B >: D2? 321331W R bRK K q < 91 >< 9C >< 9D > :G < 95 > m < 9F > < 9D > N ; < 9D > < o < 3. In class, we derived dierence approximations for the second derivative in x of the form
n n un 2u k+1 2uk + uk1 , x2 (x)2

where we have dened un k = u(k x, nt). Use a similar methodology to get a dierence approximation n n n n for the fourth derivative in terms of un k+2 , uk+1 , uk , uk1 , uk2 . What is the size of the error in your approximation? Solution: We write u 4(x)2 2 u 8(x)3 3 u 16(x)4 4 u + + O((x)5 ) + x 2 x2 6 x3 24 x4 u (x)2 2 u (x)3 3 u (x)4 4 u n + + O((x)5 ) + un k1 = uk (x) x 2 x2 6 x3 24 x4 n un k = uk
n un k2 = uk 2(x) n un k+1 = uk + (x)

un k+2

u (x)2 2 u (x)3 3 u (x)4 4 u + + + O((x)5 ) + x 2 x2 6 x3 24 x4 u 4(x)2 2 u 8(x)3 3 u 16(x)4 4 u + + + O((x)5 ) + = un k + 2(x) x 2 x2 6 x3 24 x4

So we want to nd 2 , 1 , 0 , 1 , 2 so that if we consider the linear combination

2

i un k +i ,
i=2

then all terms before the fourth order cancel. This gives 2 + 1 + 0 + 1 + 2 = 0, 22 1 + 1 + 22 = 0, 42 + 1 + 1 + 42 = 0, 82 1 + 1 + 82 = 0. Of course this system is underdetermined, but we nd that 1 = 42 , 0 = 62 , 1 = 1 , 2 = 2 . Moreover, if we actually take that linear combination, we nd that we obtain 2 or 2 1 1 2 + 3 6 6 3 2 (x)4 (x)4 4u + O((x)5 ) x4

and therefore we choose 2

4u + O((x)5 ), x4 = 1/(x)4 . This makes the error O(x).

4. Consider the heat equation ut = uxx dened on x [0, 5], t > 0 with initial condition u(x, 0) = x(5 x) and boundary conditions u(0, t) = u(5, t) = 0. (a) Use the discretization scheme we dened in class (forward dierence in time, second centered dierence in space) with x = 1, t = 1/4. Compute two time steps forward (i.e. compute the solution at t = 1/2). (b) Do the same, except now choose t = 1/8. Compute forward four steps, again computing until time t = 1/2. (c) Now set x = 1/2, and return t to 1/4. Compute two steps forward. (d) Compare all of the answers obtained above; explain your observations. 2

Solution: In all cases, our discretization will be of the form

+1 n n un = (1 2)un k + (uk+1 + uk1 ), k

where = t/(x)2 . (a) For this case we have = 1/4, so our scheme is
+1 un = k

1 n 1 n u + (u + un k1 ). 2 k 4 k+1

Our rst row is given by 0, 4, 6, 6, 4, 0. The next row up will be given by 0, 3.5, 5.5, 5.5, 3.5, 0 (notice the left and right edges are zero because of BC!), and nally the third row is 0, 3.125, 5, 5, 3.125, 0. The rest of the cases are similar, except more work. In the other two cases we have = 1/8 and = 1, respectively. You should observe that the rst two cases work reasonably, but the third case shows signs of instability. 5. (Strauss 8.2.11.) Write down a discretization scheme for ut = auxx + bu where we use forward dierence in time, and centered second dierence in space. Dene = t/(x)2 and nd the condition on for this scheme to be stable. Solution: Using the dierence schemes, we have
n +1 un 2un un un k + u k 1 k k + bun = a k+1 k, t (x)2

and solving gives

+1 n n un = (1 2a + bt)un k + a(uk+1 + uk1 ). k

As for stability, we do the standard argument, take un k = Xk Tn , and we obtain (1 2a + bt) + a X k 1 Xk+1 + Xk Xk =

and Tn+1 = Tn , so for stability we need | | < 1. We make the Ansatz Xk = (ei(x) )k , and furthermore notice that in the limit of small grids the t term disappears, and we have = (1 2a) + a(ei(x) + ei(x) ) = (1 2a) + 2a cos((x)) = 1 2a(1 cos((x))). Since the 1 cos() term is always positive, we cannot have > 1, but we could have < 1. Taking the worst-case scenario into account, i.e. 1 cos((x)) < 2, we need to specify that 1 4a < 1, or < 1 . 2a

10

Math 110B: Homework 1

April 11, 2012

All problems are from Strauss, Partial Dierential Equations: An Introduction, second edition.

6.1.13

A function u(x) is subharmonic in D if u 0 in D. Prove that its maximum value is attained on bdy D. Proof. We can complete this exercise by following the proof of the maximum principle for harmonic functions on page 155 with a slight modication. Fix > 0 and dene v (x) = u(x) + |x|2 . Computing in two dimensions as in the text, v = u + (x2 + y 2 ) 0 + 4 > 0, the only dierence being the rst sign because our function is subharmonic instead of harmonic. From here, we follow the proof in the text exactly to prove the result.

6.2.1

Solve uxx + uyy = 0 in the rectangle 0 < x < a, 0 < y < b with the following boundary conditions: ux = a on x = 0, ux = 0 on x = a, uy = b on y = 0, and uy = 0 on y = b. Proof. The hint that suggests guessing that the solution is a quadratic polynomial in x and y , i.e. u(x, y ) = Ax2 + Bxy + Cy 2 + Dx + Ey + F. We have the partial derivatives ux (x, y ) = 2Ax + By + D uy (x, y ) = Bx + 2Cy + E. Now we use the boundary conditions to solve for the constants. Since a = ux (0, y ) = By + D must hold for all values of y , we have B = 0 and D = a. Using 0 = ux (a, y ) = 2Aa + By + D = 2Aa a,

we nd that A = 1/2. Similarly b = uy (x, 0) = E . Solving 0 = uy (x, b) = 2Cb + b gives B = 1/2. Hence 1 1 u(x, y ) = x2 y 2 ax + by + F 2 2 solves the equation for any constant F .

6.2.3

Find the harmonic function u(x, y ) in the square {0 < x < , 0 < y < } with the boundary conditions uy = 0 for y = 0 and for y = , u = 0 for x = 0, and u = cos2 y for x = . Proof. We separate variables u(x, y ) = X (x)Y (y ), and so X /X + Y /Y = 0. Hence there exists a constant such that Y + Y = 0 on 0 < y < and X X = 0 on 0 < x < . We chose as such since the boundary conditions are homogeneous in y . This way, will be a non-negative eigenvalue and we will be able to nd the eigenvalues and eigenfunctions as in the previous chapter. Y + Y = 0 has solutions of the form Y (y ) = Cy + D if = 0 and Y (y ) = A cos(y ) + B sin(x) where 2 = if > 0. If = 0, Y solves the ODE Y = 0. Integrating twice gives Y (y ) = Cy + D. Using the boundary condition for y , 0 = Y (0) = C shows that the rst eigenvalue is 0 with the constant eigenfunction Y (y ) = D . If > 0, 0 = Y (0) = A sin(0) + B cos(0) = B and hence B = 0. The second boundary condition 0 = Y ( ) = A sin( ) implies that = 1, 2, 3, . . .. Hence we have eigenvalues n = n2 with eigenfunctions Yn (y ) = cos(nx) for n = 0, 1, 2, . . . . Now we solve the ODE for X . When n = 0, we have X = 0 and so X (x) = Cx + D. Using the boundary condition 0 = X (0) = D, X (x) = Cx. When n > 0, X n X = 0 has solutions of the form A cosh(nx) + B sinh(nx). The boundary condition 0 = X (0) = A implies that X (x) = B sinh(nx). Putting together our solutions for x and y , we have the series solution

u(x, y ) = A0 x +
n=1

An sinh(nx) cos(ny ).

Substituting in the nal boundary condition, 1 1 + cos(2y ) = u(, y ) = A0 + 2 2

An sinh(n ) cos(ny ).
n=1

By matching coecients, A0 = 1 , 2 A2 = 1 , 2 sinh(2 )

and An = 0 for n = 1 and n 3. This gives the solution u(x, y ) = x 1 + sinh(nx) cos(ny ). 2 2 sinh(2 )

6.2.4

Find the harmonic function in the square {0 < x < 1, 0 < y < 1} with the boundary conditions u(x, 0) = x, u(x, 1) = 0, ux (0, y ) = 0, ux (1, y ) = y 2 . Proof. Attempting to solve this using separation of variables wont quite work since we will not be able to nd eigenfunctions without homogeneous boundary conditions in either variable. Instead, we will break the problem in to two problems which each have homogeneous boundary conditions in one variable. Notice that if u1 is a harmonic function on the unit square with boundary conditions u1 (x, 0) = x, u1 (x, 1) = 0, (u1 )x (0, y ) = 0, (u1 )x (1, y ) = 0 and u2 is a harmonic function on the unit square with boundary conditions u2 (x, 0) = 0, u2 (x, 1) = 0, (u2 )x (0, y ) = 0, (u2 )x (1, y ) = y 2 , then u = u1 + u2 solves the initial problem. Applying separation of variables to u1 (x, y ) = X (x)Y (y ) gives X /X + Y /Y . Since u1 has homogeneous boundary conditions in the x variable, we choose the sign of so that X + X = 0 and Y Y = 0. Solving the ODE for X as in the previous problem, we nd eigenvalues n = (n )2 and eigenfunctions Xn (x) = cos(nx). If = 0, then Y = 0 has the solution Y (y ) = Cy + D. Using the boundary condition 0 = Y (1) = C + D, we nd that C = D and hence Y (y ) = D(1 y ). If > 0, then Y Y = 0 has solutions of the form Y (y ) = A cosh(ny ) + B sinh(ny ). The boundary condition implies that 0 = Y (1) = A cosh(n )+B sinh(n ), and so B = A cotanh(n ), giving the solution Y (y ) = A (cosh(ny ) cotanh(n ) sinh(ny )) . Putting these together, 1 u1 (x, y ) = A0 (1 y ) + 2

An cosh(ny ) cotanh(n ) sinh(ny ) cos(nx).

n=1

(1)

Using the remaining boundary condition, 1 x = u1 (x, 0) = A0 + 2 which is just the Fourier cosine series of x. Hence
1

An cos(nx),
n=1

A0 = 2
0

x dx = 1 3

and for n = 1, 2, 3, . . .,
1

An = 2
0

x cos(nx) dx =

2 ((1)n 1) . (n )2

Now we solve the boundary problem for u2 . We separate variables and choose the sign of so that Y + Y = 0 and X X = 0. Solving the homogeneous problem for y , if = 0 we have the solution Y (y ) = Cy + D, but the boundary conditions imply that C = D = 0 hence 0 is not an eigenvalue. For > 0, we have solutions of the form Y (y ) = A cos(y ) + B sin(y ). The boundary condition 0 = Y (0) = A cos(0) + B sin(0) = A implies that A = 0. The second boundary condition 0 = Y (1) = B sin( ) implies that = , 2, 3, . . .. Hence we have eigenvalues n = (n )2 with eigenfunctions Yn (y ) = sin(ny ) for n = 1, 2, 3, . . .. Solving the ODE X n X = 0 gives X (x) = A cosh(nx) + B sinh(nx). The boundary condition 0 = X (0) = An sinh(0) + Bn cosh(0) = Bn implies that B = 0. Hence X (x) = A cosh(nx). Putting X and Y together gives

u2 (x, y ) =
n=1

Bn cosh(nx) sin(ny ).

(2)

The nal boundary condition gives the equation

y = (u2 )x (1, y ) =
n=1

Bn n sinh(n ) sin(ny ).

Using the formula for the Fourier sine series coecients, Bn = 2 n sinh(n )
1

y 2 sin(ny ) dy =
0

2 (1)n 2((1)n 1) + . n sinh(n ) n (n )3

Taking u = u1 + u2 by adding (1) and (2) with the coecients An and Bn as above completes the problem.

Math 110B: Homework 2

April 23, 2012

All problems are from Strauss, Partial Dierential Equations: An Introduction, second edition.

6.3.1

Suppose that u is a harmonic function in the disk D = {r < 2} and that u = 3 sin 2 + 1 for r = 2. Without nding the solutions, answer the following questions: 1. Find the maximum value of u in D. 2. Calculate the value of u at the origin. Proof. For part 1, by the maximum principle the maximum must occur on the boundary of D, which is the set {r = 2}. So, max u = max u = max 3 sin 2 + 1 = 3 + 1 = 4.
D r=2

(1)

We calculate part 2 using the mean value property taking a = 2: u(0) = a2 2a u(x ) 1 ds = a2 2
2

u() d =
0

x =a

1 2

3 sin 2 + 1 d = 1.
0

(2)

6.3.2

Solve uxx + uyy = 0 in the disk {r < a} with the boundary condition u = 1 + 3 sin on r = a. Proof. Using equation (10) on page 167, u has the form 1 u(r, ) = A0 + 2

rn (An cos n + Bn sin n) .

n=1

(3)

At r = a, 1 1 + 3 sin = u(a, ) = A0 + 2

an (An cos n + Bn sin n) .

n=1

(4)

Comparing terms, we have A0 = 2, B1 = 3/a and all other coecients are zero. So u(r) = 1 + 3r sin . a (5)

6.3.3
3 4

Same for the boundary condition u = sin3 ]theta = Proof. As in 6.3.2, at r = a,

sin 1 4 sin 3 .

3 1 1 sin sin 3 = u(a, ) = A0 + 4 4 2 Comparing terms, we have B1 =

3 4a ,

an (An cos n + Bn sin n) .

n=1

(6)

1 B3 = 4a 3 and all other coecients are zero. So

u(r, ) =

3r r3 sin 3 sin 3. 4a 4a

(7)

6.3.4

Show that P (r, ) is a harmonic function in D by using polar coordinates. That is, use (6.1.5) on the rst expression in (17). Proof. Observe that = 2 1 1 2 + + , r2 r r r2 2
2P , r2

(8) and
2P 2

in polar coordinates. This question amounts to calculating P r , that P = 0. This is a long computation, but you can do it.

and then showing

6.4.1

Solve uxx + uyy = 0 in the exterior r > a of a disk, with the boundary condition u = 1 + 3 sin on r = a, and the condition at innity that u be bounded as r . Proof. As in the example 3 in chapter 6.4, the solution is of the form 1 u(r, ) = A0 + 2

rn (An cos n + Bn sin n) .

n=1

(9)

At r = a, 1 1 + 3 sin = u(a, ) = A0 + 2

an (An cos n + Bn sin n) .

n=1

(10)

Comparing terms, we have A0 = 2, B1 = 3a and all other coecients are zero. So u(r, ) = 1 + 3a sin . r (11)

6.4.2

Solve uxx + uyy = 0 in the disk r < a with boundary condition u hu = f () r where f () is an arbitrary function. Write the answer in terms of the Fourier coecients of f (). Proof. The solution is of the form 1 u(r, ) = A0 + 2

rn (An cos n + Bn sin n) .

n=1

(12)

by equation (10) on page 167. Dierentiating, h u(r, ) hu(r, ) = A0 + r 2 At r = a, h f () = A0 + 2

nrn1 hrn (An cos n + Bn sin n) .

n=1

(13)

nan1 han (An cos n + Bn sin n) .

n=1

(14)

Using the formula for the Fourier coecients, 2 2 2 Bn = 2 An = provided that h = n/a.
2 0 2 0

nan1

f () cos n d han f () sin n d nan1 han

(15) (16)

6.4.4

Derive Poissons formula (9) for the exterior of a circle.

Proof. As in the example 3 in chapter 6.4, the solution for the exterior of a circle with boundary condition u(a, ) = h() and u bounded as r is of the form 1 u(r, ) = A0 + 2 with coecients an 2 h() cos n d 0 an 2 Bn = h() sin n d. 0 Now we calculate exactly as on page 167 for Poissons formula for the interior of a circle: An = u(r, ) = 1 2 1 2 1 2 1 2
2

rn (An cos n + Bn sin n)

n=1

(17)

(18) (19)

h() d +
0 2 0 2 n=1

an rn a r a r
n

(cos n cos n + sin n sin n) d

0 n

h() 1 + 2
n=1

cos n ( )

d a r
n

h() 1 +
0 2 n=1

ein() +
n=1

ein()

= =

h() 1 +
0 2 0

aei() aei() + r aei() r aei()

d (20)

r2 a2 2

a2

h() d. 2ar cos( ) + r2

6.4.6

Find the harmonic function u in the semidisk {r < 1, 0 < < } with u vanishing on the diameter ( = 0, ) and u = sin sin 2 on r = 1. Proof. Using equation (5) on page 173, the solution is of the form

u(r, ) =
n=1

An rn sin n.

(21)

At r = 1, sin sin 2 = u(1, ) =

An sin n.
n=1

(22)

Comparing terms gives A1 = , A2 = 1 and all other coecients are zero. So u(r, ) = r sin r2 sin 2. (23)

6.4.9

Solve uxx + uyy = 0 in the wedge r < a, 0 < < with the boundary conditions u = on r = a, u = 0 on = 0, and u = on = . Proof. The hint suggests that we guess a solution independent of r. We can quickly see that u(r, ) satises all of the boundary conditions. Applying the Laplace operator in polar coordinates = we see that u = 0, hence u(r, ) = solves the boundary value problem. (25) 2 1 1 2 + + 2 2, 2 r r r r (24)

Math 110B: Homework 3

May 6, 2012

All problems are from Strauss, Partial Dierential Equations: An Introduction, second edition.

7.1.1

Derive the three-dimensional maximum principle from the mean value property. Proof. We need to show that if u is a harmonic function on D R3 , then any maximum occurs on the boundary of D. Assume that a maximum occurs at a point x0 in the interior of D. By translation, we may assume that x0 = 0. Take any ball of radius a so that B (0, a) D. By the mean value property, u(0) = ua where ua is the average value of u on the circle of radius a. Since the left hand side of the above equation is independent of a, we nd that u(x) = u(0) for all x contained in B (0, a) such that the ball is contained in D. Repeating this argument for enough points so that the balls cover D shows that u is constant on all of D completing the proof.

7.1.3
u n

Prove the uniqueness of the Robin problem boundary.

+ a(x)u(x) = h(x) provided that a(x) > 0 on the

Proof. Let u1 and u2 solve the Robin problem for harmonic functions u = 0 on D u + a(x)u(x) = h(x) on D. n Then the function w = u1 u2 solves the homogeneous problem w = 0 on D w + a(x)w(x) = 0 on D. n Recall the that the Dirichlet energy is dened for a function u as E (u) =
D

|u|2 dV 0. 1

Using Greens rst identity and the fact that w = 0, E (w) =

D

u dS = n

au2 dS 0.
D

Combining these two inequalities, we have E (w) = 0 and hence w = 0. This implies that w = C for some constant C , but the boundary conditions for w, 0= imply that w = C = 0 and so u1 = u2 . w + a(x)w(x) = a(xC n

7.1.5

Prove Dirichlets principle for the Neumann boundary condition. It asserts that among all realvalued functions w(x) on D the quantity E (w) = 1 2 |w|2 dV
D D

hw dS

is the smallest for w = u where u is the solution to the Neumann problem u = 0 on D u = h(x). on D n It is required to assume that the average of the given function h(x) is zero. Proof. Let w be any other smooth function and v = u w and let M > 0 be such that u(x) M for all x in D. Then E (w) = 1 2 |u v |2 dV
D D

hu hv dS u v dV +
D D

= E (u) +
D

|v |2 dV v
D

hv dS

E (u) + 0 = E (u)

u + n

hv dS
D

where on the second to last line we used the fact that E (v ) 0 and Greens rst identity.

7.2.1
1 2 u (log |x x0 |) log |x x0 | dS n n

Derive the representation formula for harmonic functions u(x0 ) = in two dimenstions. 2 u(x)
D

Proof. The proof proceeds exactly as in the text on page 186, except we take 1 log |x x0 |. 2 Let u be a harmonic function. Notice that v is not dened at x0 , so we will consider the domain D = D \ B (x0 , ). We translate so that x0 is the origin, thus x = |x x0 | = r, v ( x) = v (r) = and 1 log r 2

1 1 2 1 1 2 + + v= + +0=0 r2 r r r2 2 2r2 2r2 on D . Now we apply Greens second identity to u and v on D . The left hand side is 0, and so v = 0= 1 2 u
D

u 1 (log r) log r dS + n n 2

u
B (0,)

u (log r) log r dS. n n

Note that the integral over D is what we want, and so we will evaluate the integral over B (0, ). The outward facing normal for D points in the negative r direction (see Figure 1 on page 186 in the text), so we calculate 1 2 u
B (0,)

(log r) = n

B (0)

u 1 dS = 2r 2

udS = u
B (0,)

where u is the average value of u on the circle of radius . 1 2 where u log log rdS = n 2
u n

B (0,)

B (0,)

u 1 dS = log n 2

B (0,)

u 1 u dS = log n 2 n

u n

is the average value of 1 0 2 lim u

B (0,)

on the circle of radius . Then taking 0,

u u (log r) log r dS = u(0) 0 |r=0 = u(0) n n n

completing the proof.

7.2.2

Let (x) be any C 2 function dened on all of three dimensional space that vanishes outside some sphere. Show that 1 1 (0) = (x)dV 4 D | x| where D is the region where is not zero. Proof. We follow the calculation on page 186 with u = , x0 = with the same choice of v . We apply Greens second identity to and v on D = D \ B (0, ) (where v = 0): 1 4
D

1 1 (x)dV = |x| 4

1 1 1 + dS + n |x| |x| n 4

B (0,)

1 1 + dS. n |x| |x| n

Since is C 2 and vanishes outside of D, the rst integral on the right hand side is zero. Now evaluting the right hand side and taking 0 gives the result. 3

7.2.3

Give yet another derivation of the mean value property in three dimensions by choosing D to be a ball and x0 its center in the representation formula. Proof. Let u = 0 and apply the representation formula with D = B (x0 , a): u(x0 ) = 1 4 u
B (x0 ,a)

1 1 dS + n |x x0 | 4

B (x0 ,a)

1 u dS. |x x0 | n

In polar coordinates on B (x0 , a), 1 1 1 = = 2, n |x x0 | r r r so the rst integral on the right hand side becomes 1 4 u
B (x0 ,a)

1 1 dS = n |x x0 | 4a2

u dS.
B (x0 ,a)

Using equation (2) on page 180, the second integral is 1 4 and so u(x0 ) = 1 1 u dS = |x x0 | n 4a2 1 4a2 u 1 dS = n 4a2 u dS = 0,
B (x0 ,a)

B (x0 ,a)

B (x0 ,a)

u dS.
B (x0 ,a)

Math 110B: Homework 4

May 6, 2012

All problems are from Strauss, Partial Dierential Equations: An Introduction, second edition.

7.3.1

Show that the Greens function is unique. Proof. Let G1 and G2 be Greens functions for the operator and the domain D at the point x0 . Let H1 = G1 + 1/(4 |x x0 |) and H2 = G2 + 1/(4 |x x0 |). By properties (i) and (iii), G1 G2 = H1 H2 has continuous second derivatives everywhere in D and is harmonic on all of D. By property (ii), G1 G2 solves the boundary value problem u = 0 in D, u = 0 on D.

We observe that u = 0 solves the above boundary value problem. But by the uniqueness of solutions of the Dirichlet problem, we must have that G1 G2 = 0.

7.3.2

Prove the following theorem: the solution of the problem u = f in D, is given by u(x0 ) =
D

u = h on D

h(x)

G(x, x0 ) dS + n

f (x)G(x, x0 ) dV.
D

Proof. Following the proof of the representation formula (the only dierence is u = f instead of zero), with 1 v ( x) = 4 |x x0 | we have u(x0 ) =
D

v u v dS + n n

f v dV.
D

Then H (x, x0 ) = G(x, x0 ) v (x) is harmonic on D, so by Greens second identity, 0=

D

H u H dS. n n 1

Adding the above two equations, we have u(x0 ) = = G(x, x0 ) u G( x, x0 ) dS + f v dV n n D D G(x, x0 ) h(x) dS + f (x)G(x, x0 ) dV n D D u

since G vanishes on the boundary of D.

7.4.1

Find the one-dimensional Greens function for the interval (0, l). The three properties dening it can be restated as follows. (i) It solves G (x) = 0 for x = x0 (harmonic). (ii) G(0) = G(l) = 0. (iii) G(x) is continuous at x0 and G(x) + 1 2 |x x0 | is harmonic at x0 . Proof. Let H (x) = G(x) + 1 2 |x x0 |. Then since H is harmonic by (iii), H ( x ) = A1 x + A0 . Using (ii), 1 1 A0 = H (0) = G(0) + |0 x0 | = x0 . 2 2 1 1 1 A1 l + x0 = H (l) = G(l) + |l x0 | = (l x0 ) 2 2 2 hence A1 = Then 1 G(x) = H (x) |x x0 | = 2 x(l x0 )/l : 0 x x0 l x0 (l x)/l : 0 x0 x l 1 x0 . 2 l

7.4.2

Verify directly from (3) or (4) that the solution of the half-space problem satises the condition at innity: u(x) 0 as |x| . Assume that h(x, y ) is a continuous function that vanishes outside some circle.

Proof. Since h(x, y ) is continuous and vanishes outside of some circle it is bounded, i.e. there is a constant M > 0 such that h M . Clearly, z0 |x0 |. By the triangle inequality, |x0 | |x| + |x x0 |. Rearranging, 1 1 . |x x0 |3 (|x| |x0 |)3 Now we substitute all of this into (4) and take the limit as |x0 | :
|x0 |

lim u(x0 ) =

z0 h(x) dS 2 | x x0 |3 D | x0 | M lim dS 3 |x0 | 2 D (|x| |x0 |) |x0 | M lim dS = 2 ( | x | |x0 |)3 D |x0 | = 0.
|x0 |

lim

7.4.3

Show directly from (3) that the boundary condition is satised: u(x0 , y0 , z0 ) h(x0 , y0 ) as z0 0. Assume h(x, y ) is continuous and bounded. Proof. We use a variant of cylindrical coordinates, x = x0 + sz0 cos y = y0 + sz0 sin . Then
2 s 2 z0 = (x x0 )2 + (y y0 )2

and since det x/s x/ y/s y/ = det z0 cos sz0 sin z0 sin sz0 cos
2 = sz0 ,

2 dxdy = sz0 dsd.

Now substituting all of this into (3) from the text and taking z0 0,
z0 0

lim u(x0 , y0 , z0 ) =

z0 2 3/2 [(x x0 )2 + (y y0 )2 + z0 ] h(x, y ) dxdy z0 0 2 2 1 = lim s(s2 + 1)3/2 h(x0 + sz0 cos , y0 + sz0 sin ) drd z0 0 2 0 0 2 1 = s(s2 + 1)3/2 h(x0 , y0 ) drd 2 0 0 = h(x0 , y0 ). lim

7.4.5

Notice that the function xy is harmonic in the half-plane {y < 0} and vanishes on the boundary line {y = 0}. The function 0 has the same properties. Does this mean that the solution is not unique? Explain. Proof. In fact, any function of the form (Ax + B )y has the above properties. However, uniqueness requires that the function is bounded at innity, i.e. u(x, y ) 0 as |x0 | . The only solution satisfying this extra condition is the zero function.

7.4.6

(a) Find the Greens function in the half-plane {y > 0}. Proof. Following the rst section of 7.4, 1 1 G( x, x0 ) = log |x x0 | log |x x0 |. 2 2

(b) Use it to solve the Dirichlet problem in the half-plane with boundary values h(x). Proof. Using our representation formula in two dimensions from Exercise 7.2.1, and following as in the proof of Theorem 7.3.1, u(x0 , y0 ) =
{y>0}

u(x, y )

G(x, x0 ) dx. n

The boundary of the half-plane is the set y = 0, so the outward normal is in the negative y direction. Hence u(x0 , y0 ) = u(x, 0) G(x, y, x0 , y0 ) dx y y=0 1 y0 = h(x) 2 dx. y0 + (x x0 )2

(c) Calculate the solution with u(x, 0) = 1. Proof. We use the substitution w = x x0 . u(x0 , y0 ) = 1 y0 2 + (x x )2 dx y0 0 y0 1 = 2 + w 2 dw y0 1 = arctan(w/y0 ) = = 1.

7.4.7

(a) If u(x, y ) = f (x/y ) is a harmonic function, solve the ODE satised by f . Proof. 0 = u = 2 2 + x2 y 2 f (x/y ) = 1 2x x2 f ( x/y ) + f ( x/y ) + f (x/y ). y2 y3 y4

Now we multiply the equation by y 2 and make the substitution z = x/y to obtain 0 = f (z )(1 + z 2 ) + 2zf (z ) = ((1 + z 2 )f (z )) . Integrating once, A = (1 + z 2 )f (z ). Solving for f (z ) and integrating again, f (z ) = A arctan(z ) + B.

(b) Show that u/r 0. Proof. Recall that in polar coordinates, x = r cos and x/r = cos = x/r and similarly y/r = y/r. Then by the chain rule, u u x u y 1 x x y = + = f (x/y ) 2 f (x/y ) = 0. r x r y r y r y r

(c) Suppose that v (x, y ) is any function in {y > 0} such that v/r 0. Show that v (x, y ) is a function of the quotient x/y . Proof. Notice that if v/r 0, then v only depends on = arctan(x/y ) and so v is a function of x/y . (d) Find the boundary values limy0 u(x, y ) = h(x). Proof. For x < 0,
y 0

lim u(x, y ) = lim A arctan(x/y ) + B = A

y 0

+ B. 2

For x = 0,
y 0

lim u(0, y ) = lim A arctan(0) + B = B.

y 0

For x > 0,
y 0

lim u(x, y ) = lim A arctan(x/y ) + B = A

y 0

+ B. 2

(e) Show that your answer to parts (c) and (d) agrees with the general formula from Exercise 6. Proof. We let h(x) = in our formula from 7.4.6(b): u(x0 , y0 ) = y0 (A/2 + B ) 1 y0 (A/2 + B ) dx + 2 2 + (x x )2 dx 2 0 y0 0 y0 + (x x0 ) 0 x x0 A B x x0 A B arctan + arctan = + + 2 y0 2 y0 0 A B x0 A B x0 = + arctan + + + arctan 2 y0 2 2 2 y0 x0 = A arctan + B. y0 1
0

A 2 +B :x<0 A + B :x>0 2

7.4.21

The Neumann function N (x, y ) for a domain D is dened exactly like the Greens function in Section 7.3 except that (ii) is replaced by the Neumann boundary condition N = c for x D n for a suitable constant c. (a) Show that c = 1/A, where A is the area of D. Proof. For simplicity, we will translate so that x0 = 0. Let D = D \ B (0, ). Let v (x) = 1 4 |x x0 |

and H = N v . Then H is harmonic on D, and N and v are harmonic on D , so 0 = lim

0

N dV lim
D

v dV
D

= lim =

=
D

= lim

v dS n D D N v N dS + lim dS dS lim 0 0 n B (0,) n D n H N v dS + lim dS lim dS 0 0 n B (0,) n B (0,) n N v dS lim dS. 0 n n B (0,) B (0,) N dS lim 0 n 6

B (0,)

v dS n

(1)

Now we calculate 0 = lim

0

N dV
D

= lim =

N dS n
B (0,)

N dS + lim 0 n

N dS. n

(2)

Now substituting (1) into (2), 0=

D

N dS + lim 0 n

B (0,)

v dS = n

c dS + lim
D

B (0,)

v dS. n

As in the proof of the representation formula, we use spherical coordinates and observe that = r , so on B (0, ), n 0 = cA + lim
0

B (0,)

= cA lim

B (0,)

1 dS r 4r 1 dS 4r2 dS

1 42 = cA lim 1 = cA lim
0 0

B (0,)

= cA 1.

(b) State and prove the analog of Theorem 7.3.1, expressing the solution of the Neumann problem in terms of the Neumann function. Theorem 9.1. If N (x, x0 ) is the Neumann function, then the solution of the Neumann problem u = 0 on D, is given by the formula u(x0 ) =
D

u = h on D n u dS. n

uc N

Proof. Let v (x) = 1 4 |x x0 |

and H = N v . Then by the representation formula, u(x0 ) =

D

u v v dS. n n

(3)

By Greens second identity, 0=

D

H u H dS. n n

(4)

Adding (3) and (4), u(x0 ) =

D

u N N dS = n n

uc N
D

u dS. n

Math 110B: Homework 5

May 14, 2012

All problems are from Strauss, Partial Dierential Equations: An Introduction, second edition.

9.1.1

Find all three-dimensional plane waves; that is, all the solutions of the wave equation of the form u(x, t) = f (k x ct) where x is a xed vector and f is a function of one variable. Proof. By the chain rule, 0 = utt c2 u = c2 (1 |k|2 )f (k x ct). So either |k| = 1 or we have f (k x ct) = A(k x ct) + B by integrating twice.

9.1.2

Verify that (c2 t2 x2 y 2 z 2 )1 satises the wave equation except on the light cone. Proof. Away from the light cone, we calculate: ut = c2 t(c2 t2 x2 y 2 z 2 )2 utt = 8c4 t2 2c2 (c2 t2 x2 y 2 z 2 ) (c2 t2 x2 y 2 z 2 )3

ux = 2x(c2 t2 x2 y 2 z 2 )2 uxx = 8x2 + 2(c2 t2 x2 y 2 z 2 ) (c2 t2 x2 y 2 z 2 )3

and similarly for uyy and uzz . Then adding these together, utt c2 (uxx + uyy + uzz ) = 0.

9.1.3

Verify that (c2 t2 x2 y 2 )1/2 satises the two-dimensional wave equation except on the cone {x2 + y 2 = c2 t2 }. Proof. We calculate: utt = 3c4 t4 c2 (c2 t2 x2 y 2 ) (c2 t2 x2 y 2 )5/2 3x2 + (c2 t2 x2 y 2 ) (c2 t2 x2 y 2 )5/2

uxx =

and similarly for uyy . Adding these gives the result.

9.1.4

Thinking of the coordinates of space-time as 4-vectors (x, y, z, t), let be the diagonal matrix with the diagonal entries 1, 1, 1, 1. Another matrix L is called a Lorentz transformation if L has inverse and L1 = LT for LT is the transpose of L. (a) If L and M are Lorentz, show that LM and L1 also are. Proof. Note that = I so 1 = . Then (LM )1 = M 1 L1 = M T LT = (LM )T , so LM is Lorentz. Also, by rearranging the Lorentz condition and taking transpose, L = (L1 )T , so (L1 )1 = L = (L1 )T and L1 is Lorentz. (b) Show that L is Lorentz if and only if m(Lv) = m(v) for all 4-vectors v = (x, y, z, t), where m(v = x2 + y 2 + z 2 t2 is called the Lorentz metric. Proof. By rearranging the Lorentz condition, = LT L. Then m(v) = v v = (LT L)v v = (Lv) (Lv) = m(Lv).

(c) If u(x, y, z, t) is any function and L is Lorentz, let U (x, y, z, t) = U (L(x, y, z, t)). Show that Uxx + Uyy + Uzz Utt = uxx + uyy + uzz utt .

Proof. Let Lij be the entry in the i-th row and j -th column of L. We will use the notation x1 = x, x2 = y, x3 = z, x4 = t. Then by the chain rule,
4

Ui =
k=1 4

uk Lki

Uii =
k,l=1

ukl Lki Lli .

Then regrouping terms,

4 4

U11 + U22 + U33 U44 =

k,l=1

ukl (Lk1 Ll1 + Lk2 Ll2 + Lk3 Ll3 Lk4 Ll4 ) =

k,1=1

ukl kl

where kl is the entry of in the k -th row and l-th column. This completes the proof.

9.1.7

For the boundary condition u/n + bu/t = 0 with b > 0, show that the energy dened by (6) decreasing with respect to time. Note: we must also assume that u satises the wave equation. Proof. Using equation (4) from the text, and then using the divergence theorem, E = t (utt c2 u)ut dV + c2 ut u n dS ut u dS n (ut )2 dS (ut u) dV

= 0 + c2 = c2 = c2 b 0, so E decreases in time.

9.2.2

Verify that (3) is correct in the case of the example u(x, y, z, t) = t. Proof. Note that if u = t, (x) = u(x, 0) = 0 and (x = ut (x, 0) = 1. Then by Kirchos formula, u(x0 , t0 ) = 1 4c2 t0 dS =
|xx0 |=ct0

1 4c2 t2 0 = t0 . 4c2 t0

9.2.3

Solve the wave equation in three dimensions with initial data = 0 and = y . Proof. By Kirchos formula u(x0 , t0 ) = 1 4c2 t0 y dS.
|xx0 |=ct0

In spherical coordinates with r xed at ct0 , y = y0 + ct0 sin sin and dS = c2 t2 0 sin dd (we are using the convention in the text where goes from 0 to 2 and goes from 0 to ), so u(x0 , t0 ) = 1 4c2 t0
2 0 0 2 2 2 (c3 t3 0 sin sin + c t0 y0 sin ) dd = t0 y0 .

9.2.4

Solve the wave equation in three dimensions with initial data = 0 and = x2 + y 2 + z 2 . Proof. Again, we use Kirchos formula in spherical coordinates with r xed at ct0 (I recommend skipping the hint in the text): x = x0 + ct0 cos sin y = y0 + ct0 sin sin z = z0 + ct0 cos dS = c2 t2 0 sin dd. u(x0 , t0 ) = 1 4c2 t0 x2 + y 2 + z 2 dS
|xx0 |=ct0 2 2 (x2 0 + y0 0 0 2 + z0 + c2 t2 0 + 2x0 t0 cos sin + 2y0 t0 sin sin

1 = 4c2 t0

+ 2z0 t0 cos )c2 t2 0 sin dd

2 2 2 2 = (x2 0 + y0 + z0 + c t0 )ct0 .

9.2.5

Where does a three-dimensional wave have to vanish if its initial data and vanish outside of a sphere?

Proof. Suppose that and vanish outside of B (0, R). Then by Kirchos formula, the solution only depends on the initial data on the sphere {|x x0 | = ct0 }. So a three dimensional wave will vanish at points (x0 , t0 ) satisfying |x0 | > R + ct0 or ct0 > R + |x0 |.

10

9.2.9

(a) For any solution of the three-dimensional wave equation with initial data vanishing outside some sphere, show that u(x, y, z, t) = 0 for xed x, y, z and large enough t. Proof. By the previous question, if we x x0 and take t0 large so that ct0 > R + |x0 | the solution will vanish. (b) Prove that u(x, y, z, t) = O(t1 ) uniformly as t . Proof. From part (a), we assume that and vanish outside of B (0, R). Let M = max {|(x)|, | (x)|, | (x)|}.
B (0,R)

By Kirchos formula, t0 u(x0 , t0 ) = 1 4c2 (x) dS + t0

B (x0 ,ct0 )

1 t0 4c2 t0

(x) dS .
B (x0 ,ct0 )

(1)

We estimate each term individually. For the rst term in (1), 1 4c2 (x)dS
B (x0 ,ct0 )

1 4c2

| (x)| dS
B (x0 ,ct0 )B (0,R)

M dS 4c2 B (x0 ,ct0 )B (0,R) M 4R2 4c2 M R2 = . c2 For the second term in (1), we change variables to u = 1 4c2 t0 (x) dS =
B (x0 ,ct0 ) 1 ct0 (x

x0 ) so that

t0 4

(x0 + ct0 u) dS.

B (0,R)

Then after dierentiating, we see that the second term is t0 1 4 (x0 + ct0 u) dS +
B (0,1)

t0 4

(x0 + ct0 u) cu dS .
B (0,1)

(2)

Changing our variables back, the rst term in (2) is bounded by t0 4c2 t2 0 (x) dS
B (x0 ,ct0 )

1 4c2 t0

|(x) dS
B (x0 ,ct0 )B (0,R)

4M R2 4c2 t0 MR 2 c

if t0 is large enough so that t0 > R. Note that we can do this since we only need to prove the result for large values of t0 . To estimate the second term in (2), t2 0 4c2 t2 0 (x)
B (x0 ,ct0 )

x x0 1 dS ct0 4c2 4M R2 4c2 M R2 . = c2 =

M dS
B (x0 ,ct0 )B (0,R)

where we used the fact that |x x0 | = ct0 on B (x0 , ct0 ). Putting all of this back into (1), we have 2M R 2 + M R |t0 u(x0 , t0 )| c2 for t0 > R. Since the constant on the right is independent of x0 and t0 , this completes the proof.

11

9.2.10

Derive the mean value property of harmonic functions u(x, y, z ) by the following method. A harmonic function is a wave that does not depend on time, so that its mean value u(r, t) = u(r) satises (5). Deduce that u(r) = u(0). Proof. Using (5), since utt = 0, 1 0 = c2 urr + 2c2 ur . r If we dene v = ru, then v solves the ODE 0 = v + 2v. Solving by separation of variables, v (r) = A/r2 . Integrating, B u(r) = + C. r But we require that our solution is bounded as r 0, so we must have B = 0. So u(r) is a constant, and so u(r) = u(0). Then (4) gives the mean value property.