Math 121A: Sample midterm solutions

1. To find a Green function solution, consider solving the given equation for an impul-
sive input of the form f ( x ) = δ( x − x 0 ). In the regions x > x 0 and x < x 0 , solutions
of the form y = emx can be searched for, which upon substitution into the differential
equation gives
m2 − k 2 = 0
and hence m = ±k. To satisfy the boundary conditions as x → ±∞, the solution
must be of the form
Aemx for x < x 0 ,

y( x ) =
Be−mx for x > x 0 ,
where A and B are constants. To satisfy continuity at x 0 it follows that
0 0
Aemx = Be−mx
and to satisfy y0+ ( x 0 ) − y0− ( x 0 ) = 1 it follows that
0 0
− Bme−mx − Amemx = 1
and thus 0 0
e−mx emx
A=− , B=−
2m 2m
so the solution can be written as
0
e−m| x − x |
y( x ) = − .
2m
Hence the Green function solution for an arbitrary source term f ( x ) can be written
as
0
Z ∞
!
e−m| x − x |
y( x ) = − f ( x 0 )dx 0 .
−∞ 2m

2. (a) The Fourier transform is

Z ∞
1  
h̃λ (α) = exp −λx2 − ixα dx
2π −∞
Z ∞ 2 !
α2

1 iα
= exp −λ x − − dx
2π −∞ 2λ 4λ
α2 r
e− 4λ π
=
2π λ
α2
e− 4λ
= √ .
4πλ
 1.2
y = hλ ( x )
y = hµ ( x )
1
y = g( x )

0.8

0.6
y

0.4

0.2

0
-3 -2 -1 0 1 2 3
x

Figure 1: Graphs of the Gaussians considered in question 2, for the case of λ = 1 and
µ = 2.
 (b) The convolution is given by
Z ∞
(hλ ∗ hµ )( x ) = hλ (ξ )hµ ( x − ξ )dξ
−∞
Z ∞  
= exp −λξ 2 − µ( x − ξ )2 dξ
−∞
Z ∞  
= exp −(λ + µ)ξ 2 + 2µxξ − µx2 dξ
−∞
Z ∞ 2 !
µ2 x 2

µx
= exp −(λ + µ) ξ − + − µx2 dξ
−∞ ( λ + µ ) λ + µ
µ2 x 2 µ2 + λµ 2
r  
π
= exp + − x
λ+µ λ+µ λ+µ
r  
π λµ 2
= exp − x .
λ+µ λ+µ

(c) If λ = 1 and µ = 2 then

2x2
r  
π
g( x ) = exp − .
3 3

and thus g is a Gaussian that is wider than both hλ and hµ . The three curves are
plotted in Fig. 1.
(d) Using the result from part (a) it can be seen that
 2 
exp − α4 λ1 + µ1
2π h̃λ (α)h̃µ (α) = p .
2 λµ

Using the result from part (b),

 2 
λ+µ
π exp
r − α
4 λµ p
g̃(α) = p λ+µ
λ+µ 4πλµ
 2 
exp − α4 λ1 + µ1
= p
2 λµ

and thus the two expressions are equal as expected.

3. (a) Taking the Fourier transform in x and the Laplace transform in t gives
Z ∞ Z ∞
1
F̃ (α, p) = dx dt f ( x, t)e− pt−ixα .
2π −∞ 0
 Taking the transform of the equation f t + c f x = b f xx gives

p F̃ − f˜(α, 0) + ciα F̃ = b(iα)2 F̃.

The Fourier transform of the initial condition is

Z ∞
1 1
f˜(α, 0) = δ( x )e−ixα dx =
2π −∞ 2π
and hence the transformed equation can be written as

1
( p + icα + bα2 ) F̃ = .
2π
Therefore
1
F̃ (α, p) = .
2π ( p + icα + bα2 )
Since the Laplace transform of e−qt is 1/( p + q), it follows that
2 )t
e−(icα+bα
f˜(α, t) =
2π
Taking the inverse Fourier transform gives
Z ∞
1 2 ) t +iαx
f ( x, t) = e−(icα+bα dα
2π −∞
Z ∞ 2 !
i ( x − ct) ( x − ct)2

1
= exp −bt α − − dα
2π −∞ 2bt 4bt
( x − ct)2
 
1
= √ exp − .
4πbt 4bt

(b) By making use of part (a) and translational symmetry in x, it can be seen that if
the initial condition is g( x ) = δ( x − x 0 ) then the solution is

( x − x 0 − ct)2
 
1
f ( x, t) = √ exp − .
4πbt 4bt

The Green function solution for an arbitrary initial condition g( x ) is therefore

Z ∞
( x − x 0 − ct)2
 
1 0
f ( x, t) = √ g( x ) exp − dx 0 .
4πbt −∞ 4bt
 (c) For the case of g( x ) = e−ax the solution is given by
Z ∞
( x − x 0 − ct)2
 
1
f ( x, t) = √ exp − − ax dx 0
0
4πbt −∞ 4bt
Z ∞
x 02 + 2(ct − x ) x 0 + (ct − x )2 + 4btax 0
 
1
= √ exp − dx 0
4πbt −∞ 4bt
Z ∞
x 02 + 2(ct − x + 2bta) x 0 + (ct − x )2
 
1
= √ exp − dx 0
4πbt −∞ 4bt
Z ∞
( x 0 + ct − x + 2bta)2

1
= √ exp −
4πbt −∞ 4bt
2
(ct − x + 2bta) − (ct − x ) 2 
+ dx 0
4bt
1 √ 
(ct − x )4bta + 4b2 t2 a2

= √ 4btπ exp
4πbt 4bt
= e−ax+ta(c+ba) .

It can be seen that f ( x, 0) = e−ax and hence the initial condition is satisfied. The
left hand side of the partial differential equation is

∂f ∂f
+c = a(c + ba) f − ca f = ba2 f
∂t ∂x
and the right hand side is
∂2 f
b
2
= ba2 f
∂x
so the partial differential equation is satisfied.

4. (a) The zeroth cosine term is

1 1
Z π
a0 = f ( x )dx = π = 1.
π −π π
The other cosine terms are given by

1 π
Z
an = f ( x ) cos nx dx
π −π
1 π
Z
= cos nx dx
π 0
1 sin nx π
 
=
π n 0
= 0.
 The sine terms are given by
1 π
Z
bn = f ( x ) sin nx dx
π −π
1 π
Z
= sin nx dx
π 0
1 h cos nx iπ
= −
π n 0
1
= [1 − cos nπ ]
nπ
2

nπ for n odd,
=
0 for n even.

Hence
∞
1 2 sin nx
f (x) = +∑ .
2 n=0 nπ

(b) The left hand side of the Parseval equation is

1 1
Z π Z π
| f ( x )|2 dx = dx = 1
π −π π 0

and the right hand side is

a20 ∞ ∞
1 4 1
+ ∑ ( a2n + bn2 ) = + 2 ∑ 2
.
2 n =1
2 π n=0 (2n + 1)

Equating the two gives

∞
1 4 1
= 2 ∑ 2
2 π n=0 (2n + 1)
which can be rearranged as
∞
1 π2
∑ 2
= .
n=0 (2n + 1) 8

5. First note that the Fourier transform of f (t) = e−at is

Z ∞ Z ∞
" #∞
e −( a + p ) t 1
F ( p) = e− pt dt = e−(a+ p)t dt = = .
0 0 a+p a+p
0

The Laplace transform of the equation is therefore given by

1
p2 Y ( p) − py(0) − y0 (0) + 6( pY ( p) − y(0)) + 8Y ( p) =
p+3
and hence
1
( p2 + 6p + 8)Y ( p) = ( p + 6) + .
p+3
Therefore
p+6 1
Y ( p) = +
( p + 2)( p + 4) ( p + 2)( p + 3)( p + 4)
 
1 2 1 1 1
= + + −
p + 2 ( p + 2)( p + 4) 2( p + 3) p + 2 p + 4
 
1 1 1 1 1 2 1
= + − + − +
p+2 p+2 p+4 2 p+2 p+3 p+4
5 1 1
= − −
2( p + 2) p + 3 2( p + 4)

and hence
5e−2t − 2e−3t − e−4t
y(t) = .
2