Math 648 Exam #1 - March 1, 2012

Read the problems carefully. Solve them according to the requirements and show the key
intermediate steps to receive full credits.

STUDENT NO. Solution keys

STUDENT NAME

Problem Scores
1
2
3
4
total

1
1. (25 points) Use the three methods (the increment method, -dependence method, and δ-
differentiation method) to derive the variation for the fixed point problem of the functional
Z 1
y
J[y] = p dx.
0 1 + y02

Method 1: the increment method

∆J = J[y + h] − J[y]
Z 1" #
y+h y
= p −p dx
0 1 + (y 0 + h0 )2 1 + y02

Using the Taylor expansion theorem and considering h and h0 being very small, we have
Z 1" #
y ∂ y 0 ∂ y y
∆J = p +h p +h 0p + h.o.t. − p dx
0 1 + (y 0 )2 ∂y 1 + (y 0 )2 ∂y 1 + (y 0 )2 1 + y02
Z 1" #
1 0 y y0
= hp −h dx + h.o.t.
0 1 + (y 0 )2 (1 + (y 0 )2 )3/2

From the definition, we have

" #
Z 1
1 y y 0
δJ = hp − h0 dx
0 1 + (y 0 )2 (1 + (y 0 )2 )3/2

Method 2: the -dependence method

Define Z 1 Z 1
y + h y
g() = J[y + h] − J[y] = p dx − p dx.
0 1 + (y 0 + h0 )2 0 1 + y02
Then
Z 1
dg d y + h
= p dx
d 0 d 1 + (y 0 + h0 )2
Z 1( )
h (y + h)(y 0 + h0 )h0
= − dx
[1 + (y 0 + h0 )2 ]3/2
p
0 1 + (y 0 + h0 )2

Taking  = 0 in the above equation, we get

Z 1( )
dg h y y 0 h0
δJ = (0) = − dx
1 + (y 0 )2 [1 + (y 0 )2 ]3/2
p
d 0

2
Method 3: the δ-differentiation method

Z 1
y
δJ = δp dx
0 1 + y02
Z 1" ! ! #
∂ y ∂ y
= p δy + 0 p δy 0 dx
0 ∂y 1+y 0 2 ∂y 1+y 0 2

Z 1" 0
#
1 yy
= p δy − 0 2 )3/2
δy 0 dx
0 2 (1 + y
0 1+y

3
2. (25 points) Find the extremal of the functional
Z 1 
2
J[y] = y 0 + yy 0 + y 0 + y 2 + yx2 dx
0

subject to the boundary conditions

y(0) = 0, y(1) = 1.

Solution.

2
F (x, y, y 0 ) = y 0 + yy 0 + y 0 + y 2 + yx2
Euler’s equation:
d
Fy − Fy0 = 0
dx
y 0 + 2y + x2 − (2y 0 + y + 1)0 = 0
Thus, BVP reads as (
y 00 − y = 12 x2
y(0) = 0, y(1) = 1
Solution of the differential equation: The general solution of the homogeneous equation y 00 −
y = 0 is y = C1 ex + C2 e−x . A particular solution of the nonhomogeneous equation y 00 − y =
0.5x2 is yp = Ax2 +Bx+C. Substituting this into the equation gives: A = −0.5, B = 0, C =
2A = −1. Thus, the general solution for the differential equation is
1
y = C1 ex + C2 e−x − x2 − 1
2
Using the boundary conditions we can find:
5
2− 1e e − 52
C1 = , C2 = .
e − 1e e − 1e

4
3. (25 points) Find the extremal of the functional
Z π
2 2
J[y] = (y 0 − y 2 − 2y sin(x))dx
0

when the values of y are not specified at the end points.

Solution. Euler’s equation is

−2y − 2 sin(x) − (2y 0 )0 = 0 ⇒ y 00 + y = − sin(x)

The values of y being not specified means that we have natural boundary conditions:

Fy0 (0) = Fy0 (π/2) = 0 ⇒ y 0 (0) = y 0 (π/2) = 0

The general solution of the differential equation has the form

y(x) = C1 cos(x) + C2 sin(x) + yp (x)

where
yp = Ax cos(x) + Bx sin(x)
Substituting yp into the differential equation, we get A = 0.5 and B = 0, or

1
yp = x cos(x)
2
and
1
y(x) = C1 cos(x) + C2 sin(x) + x cos(x)
2
Using the boundary conditions, we get
π 1
C1 = − , C2 = − .
4 2

5
4. (25 points) The famous sine-Gordon equation is Euler’s equation for the fixed point problem
of the functional Z Z  
1 2 1 2
J[u] = (ux ) − (uy ) + cos(u) dxdy,
R 2 2
where R is a bounded domain in the xy-plane. Derive the equation.
Solution 1: This example has two independent variables. We have
1 1
F (x, y, u, ux , uy ) = (ux )2 − (uy )2 + cos(u)
2 2
Euler’s equation (for the multiple variable case):
∂F ∂ ∂F ∂ ∂F
− − =0
∂u ∂x ∂ux ∂y ∂uy
This gives
− sin(u) − (ux )x + (uy )y = 0
or
uxx − uyy + sin(u) = 0

Solution 2:
Z Z
δJ = [ux δux − uy δuy − sin(u)δu] dxdy
Z ZR
= [(ux δu)x − (uy δu)y − uxx δu + uyy δy − sin(u)δu] dxdy
R
Z Z " # Z Z
ux δu
= ∇· dxdy + [−uxx + uyy − sin(u)] δudxdy
R −uy δu R

Using Gauss’ theorem and the fact that δu = 0 on the boundary of R, we have
Z " # Z Z
ux δu
δJ = n· dS + [−uxx + uyy − sin(u)] δudxdy
∂R −uy δu R
Z Z
= [−uxx + uyy − sin(u)] δudxdy
R

Set δJ = 0 and we get

Z Z
[−uxx + uyy − sin(u)] δudxdy = 0
R

Using the vanishing theorem, we obtain

−uxx + uyy − sin(u) = 0,

which is the famous sine-Gordon equation.