Math 4315 - PDEs

Sample Test 1 - Solutions

1. Solve the following first order PDEs by introducing an alternate coordi-

nate system (i.e. ( x, y) → (r, s))

(i ) xu x − yuy = 2u,
(ii ) yu x − uy = 1,

Solutions
(i) If us = u x xs + uy ys then choosing

xs = x, ys = −y, gives us = 2u.

Solving gives

x = a (r ) e s , y = b (r ) e −s , u = c(r )e2s ,

where a(r ), b(r ) and c(r ) are arbitrary functions of r. Eliminating s from
the first two and first and third gives
u
xy = a(r )b(r ) = A(r ), = c(r )/a(r )2 = B(r ).
x2

Noting again that A(r ) and B(r ) are arbitrary. Further, elimination of r
gives
u
2
= f ( xy) or u = x2 f ( xy).
x

(ii) If us = u x xs + uy ys then choosing

xs = y, ys = −1, gives us = 1.

Solving the second and third gives

y = − s + b (r ), u = s + c (r ),

giving the first as

x s = y = − s + b (r ).

1
Integrating gives
( b (r ) − s )2
x=− + a (r ),
2
where a(r ), b(r ) and c(r ) are arbitrary functions of r. Eliminating s from x
and y gives
y2
x+ = a(r ), or 2x + y2 = A(r ).
2
Eliminating s from y and u gives

u + y = c (r ) + b (r ) = B (r ).

Noting again that A(r ) and B(r ) are arbitrary. Further, elimination of r
gives
u + y = f (2x + y2 ) or u = −y + f (2x + y2 ).

2. Solve the following using the method of characteristics

2
(i ) yu x + xuy = xu, u( x, 0) = e− x
(ii ) xu x + 2uuy = u, u( x, 0) = x2

Solutions
(i) The characteristic equations are
dx dy du
= = .
y x xu
Solving the first pair, i.e.
dx dy
= gives c1 = x 2 − y2 .
y x

Solving the second and third, i.e.

dy du
= gives y = ln |u| − ln |c2 | or c2 = u e − y .
x xu
The solution is therefore given by

u e − y = f ( x 2 − y2 ) or u = e y f ( x 2 − y2 ).
2
Imposing the initial condition u( x, 0) = e− x gives
2
u( x, 0) = e0 f ( x2 − 02 ) = e− x =⇒ f ( x ) = e− x .

2
This gives the solution as
2 − y2 ) 2 +y− x2
u = ey e−( x = ey .

(ii) The characteristic equations are

dx dy du
= = .
x 2u u
Solving the first and third, i.e.
dx du u
= gives ln | x | = ln |u| − ln|c2 | or = c2 .
x u x

Solving the second and third, i.e.

dy du y y
= gives = u − c1 or c1 = u − .
2u u 2 2
The solution is therefore given by
u y y
= f (u − ) or u = x f (u − ).
x 2 2
Imposing the initial condition u( x, 0) = x2 gives
√
x 2 = x f ( x 2 − 0) =⇒ f ( x2 ) = x =⇒ f (x) = x.
This gives the solution as
r
y
u=x u− .
2

3. Solve the following nonlinear PDE

(i ) xu2x + uy = 1, u( x, 1) = x + 1.
(ii ) u x uy − 2xu x − 2yuy = 0, u( x, 0) = x2

Solution
(i) If F = xp2 + q − 1 then the characteristic equations are
xs = Fp = 2xp (1.4a)
ys = Fq = 1 (1.4b)
2
us = pFp + qFq = 2xp + q (1.4c)
ps = −( Fx + pFu ) = − p2 (1.4d)
qs = −( Fy + qFu ) = 0. (1.4e)

3
To these we associate the following initial condition. When s = 0, then

y = 1, x = r, u = r + 1, p = 1, q = 1 − r, (1.5)

where p is obtained from differentiating the initial condition and q from

the original equation.
From (1.4d) and (1.4e) we find that
1
= s + A (r ), q = B (r ).
p

From the boundary conditions (1.5) we find that A = 1 and B = 1 − 4 so

1
= s + 1, q = 1 − r.
p

From (1.4a) and (1.4a) we integrate giving

x = C (r )(s + 1)2 , y = s + D (r ).

The boundary conditions (1.5) gives A = r and D = 1. Thus,

x = r ( s + 1)2 , y = s + 1. (1.6)

From (1.4c) we have

us = 2xp2 + q = r + 1 ⇒
u = (r + 1)s + E(r ) = (r + 1)s + r + 1. (1.7)

Eliminate r and s from (1.6) and (1.7) gives the solution

x
u = y+ .
y

(ii) If F = pq − 2xp − 2yq then the characteristic equations are

xs = Fp = q − 2x (1.8a)
ys = Fq = p − 2y (1.8b)
us = pFp + qFq = 2pq − 2xp − 2yq = pq (1.8c)
ps = −( Fx + pFu ) = 2p (1.8d)
qs = −( Fy + qFu ) = 2q. (1.8e)

To these we associate the following initial condition. When s = 0, then

y = 0, x = r, u = r2 , p = 2r, q = 2r, (1.9)

4
where p is obtained from differentiating the initial condition and q from
the original equation. As we need p and q to find x, y and u, we focus on
these first. Solving (1.4d) and (1.4e) for p and q gives

p = a(r )e2s , q = b(r )e2s ,

and imposing the initial condition gives a(r ) = 2r and b(r ) = 2r. Thus,

p = 2re2s , q = 2re2s .

From (1.4c) we see that

us = 4r2 e4s ,
which integrates giving
u = r2 e4s + c(r ),
and the initial condition here gives c(r ) = 0. Substituting (1.10) into (1.4a)
and (1.4b) gives

xs + 2x − 2re2s = 0, ys + 2y − 2re2s = 0.

Solving yields
r 2s r 2s
x= e + d(r ) e−2s , y= e + e(r ) e−2s .
2 2
Applying the remaining initial conditions gives
r 0 r
r= e + d (r ) e0 , 0 = e0 + e (r ) e0 ,
2 2
showing that d(r ) = 2r and e(r ) = − 2r . Thus, we have the following para-
metric solutions
r ³ 2s ´ r ³ 2s ´
x= e + e−2s , y = e − e−2s , u = r2 e4s .
2 2
Eliminate r and s gives
u = ( x + y )2 .