332
max
lxllr'
When
e
=
0,
we understand that Exp(q/O;
C)
=
0,
which does not make
any contradiction in the results.
By the above choice of the Cauchy data
p(x)
we have for the Gevrey
order of the formal solution
f(t,
x)
-
EN-lfn(x)tn
n=O
5
CKN(N!)"ltlN,
Vt
E
S'\{O},
(8)
Remark
1.
Our Cauchy problem (CP) is only for a simple equation, but
the generalization is possible for an equation in two variables which has a
quasi homogeneity in derivatives
at
and
a,.
In fact, it is enough to follow
the arguments? consulting the arguments developed in this paper.
2.
a(0)-Gevrey summability (Review)
We, first, give a short review of the definition of the Gevrey summability
and the Gevrey sum.
Let
6'
E
R,
a
>
0
and
0
<
p
5
+m.
Then a sector
S(O,a,p)
in the
complex
t
plane is defined by
a
s(e,
a,
p)
=
{t
E
c
;
o
<
~tl
<
p,
larg(t)
-
el
<
2}
.
We refer
8,
a
and
p
as the direction, opening angle and the radius of the
sector, respectively.
Let
f(t,
x)
=
fn(x)tn
E
0,[[t]lU
with
a
>
0.
Then, by the defini-
tion, we say that
f(t,x)
is a-Gevrey summable in
0
direction (in t-plane)
if there exists an analytic function
f(t,
x)
on
S(t9,
a7r
+
E,
p)
x
{
1x1
5
r}
for
some positive constants
E,
p
and
r
such that for any closed sector
S'
with
a
vertex at the origin such that
(S'\{O})
c
S(0,
m++,
p)
there exist positive
constants
r',
C
and
K
such that the following asymptotic estimates hold
for all
N
2
1.
This kind of asymptotic expansion is called a-Gevrey asymp-
totic expansion, and it is written by
f(t,x)
Nu
f^(t,x)
in s(e,a7r+E).
(9)
In this expression, we omit to write the radius
p
of the sector
S(0,
m+~,
p),
since the radius is not an essential problem for the Gevrey summability. We
note that thanks of the condition that the opening angle of the sector is
larger than
a7r
for the definite domain of
f(t,
x),
one can show that such an
analytic function
f(t,
x)
is unique if such analytic functions exist, and it is