MATH 220: FIRST ORDER SCALAR QUASILINEAR
EQUATIONS
We now consider the quasilinear equations; these have the form
(1)
a
(
x, y, u
)
u
x
+
b
(
x, y, u
)
u
y
=
c
(
x, y, u
)
,
with
a, b, c
at least
C
1
, given real valued functions. There is an immediate diFerence
between semilinear and quasilinear equations at this point: since
a
and
b
depend
on
u
, we cannot associate a vector ±eld on
R
2
to the equation: we need to work on
R
3
at least to account for the (
x, y, u
) dependence.
To achieve this, we proceed as follows. We consider the graph
S
of
u
in
R
2
×
R
,
given by
z
=
u
(
x, y
). If we know the solution
u
, we know the graph, and conversely
we can recover
u
if we ±nd its graph, a surface in
R
3
. So let
U
(
x, y, z
) =
u
(
x, y
)
−
z
;
we need to ±nd the 0set of
U
. Now, substituting
U
into the left hand side of the
PDE we get
(
a
(
x, y, u
(
x, y
))
∂
x
+
b
(
x, y, u
(
x, y
))
∂
y
)
U
(
x, y, z
)
=
a
(
x, y, u
(
x, y
))
u
x
(
x, y
) +
b
(
x, y, u
(
x, y
))
u
y
(
x, y
)
.
If the PDE holds, this equals
c
(
x, y, u
(
x, y
)), which is in turn equal to
−
c
(
x, y, u
(
x, y
))
∂
z
U
as
∂
z
U
≡ −
1! Thus, our PDE implies that
(
a
(
x, y, u
(
x, y
))
∂
x
+
b
(
x, y, u
(
x, y
))
∂
y
+
c
(
x, y, u
(
x, y
))
∂
z
)
U
(
x, y, z
) = 0
for all
x, y, z
. This equation states that the directional derivative of
U
along the
vector ±eld (
a, b, c
) vanishes. Note that
∂
z
U
=
−
1, i.e. never vanishes, so in fact at
each point (
x, y, z
) the directional derivative can only vanish along a plane (and not
in every direction). But along any level set of
U
its directional derivative certainly
vanishes, so applied to the 0level set, where
u
(
x, y
) =
z
, the PDE states that the
vector ±eld
W
(
x, y, z
) = (
a
(
x, y, z
)
, b
(
x, y, z
)
, c
(
x, y, z
))
is tangent to
S
. We can now rephrase our problem as follows: we want to ±nd
S
as a union of integral curves of
W
. As before, integral curves of
W
, where we
introduce a parameter
r
again, satisfy
∂x
∂s
(
r, s
) =
a
(
x
(
r, s
)
, y
(
r, s
)
, z
(
r, s
))
,
∂y
∂s
(
r, s
) =
b
(
x
(
r, s
)
, y
(
r, s
)
, z
(
r, s
))
,
∂z
∂s
(
r, s
) =
c
(
x
(
r, s
)
, y
(
r, s
)
, z
(
r, s
));
these are called the characteristic ODEs and integral curves are called the
charac
teristics
, though perhaps it would be most appropriate to call them
lifted character
istics
in view of our terminology for semilinear equations. Note that although the
integral curves are in
R
3
, hence there is a twodimensional family of them, we only
care about the ones which will give
S
as their union, i.e. a onedimensional family
1
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document2
of integral curves. Our initial condition will again come from the initial condition
of the PDE, which is of the form
u
(Γ(
r
)) =
φ
(
r
)
along a curve Γ(
r
) = (Γ
1
(
r
)
,
Γ
2
(
r
)). The initial condition states that over the curve
Γ,
S
is given by
z
=
φ
(
r
), i.e. that
S
goes through the curve
˜
Γ(
r
) = (Γ
1
(
r
)
,
Γ
2
(
r
)
, φ
(
r
))
.
Our initial condition for the characteristic ODEs is then
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '10
 Vasy
 Vector Space, Constant of integration, initial condition, R R, Tangent space

Click to edit the document details