VIII.4 On the Unique Solvability of the Oseen Initial-Value Problem 529
Since λ is an arbitrary positive number and C
1
, C
2
do not depend on it, from
(VIII.4.11) and (VIII.4.12) we recover (VIII.4.4) and (VIII.4.6). Moreover,
recalling that G ∈ S(R
4
) (see Section VII.4), we prove immediately that u, Φ
are smooth f unctions that satisfy (VIII.4 .8)
1,2
. Summarizing a ll the above, we
thus obtain the existence of a smooth solution (w, φ) to (VIII.4.1 )
1,2
in the
class L
q
(A
T
), with φ ∈ L
q,3q/(3−q)
(R
3
T
) if q ∈ (1, 3), and satisfying (VIII.4.4),
and (VIII.4.6) as well, if q ∈ (1, 3). We shall next ana lyze the way in which
u (and hence w) tends to zero as t → 0
+
. To this end, we observe that from
(VIII.4.9) and (VIII.4.10), we have
u
j
(x, t) =
1
(2π)
2
Z
R
4
e
ix·ξ+it ξ
0
δ
ij
ξ
2
− ξ
j
ξ
k
(i ξ
0
+ ξ
2
+ λ)ξ
2
G
j
(ξ, ξ
0
)dξdξ
0
. (VIII.4.13)
Now, in view of the assumptions on F , it follows that G is analytic i n ξ
0
.
Therefore, by a simple application of the Cauchy integral theorem (MacRobert
1966, §27), we can show that the value of the integral
Z
∞
−∞
e
it ξ
0
G
j
(ξ, ξ
0
)
(i ξ
0
+ ξ
2
+ λ)
dξ
0
, t ∈ R
3
,
does not change (a nd therefore the value of the integral i n (VIII.4.1 3) does
not change) if we perform the integration on a line {α
0
} in the complex plane
parallel to {ξ
0
}, provided i ξ
0
+ ξ
2
+ λ does not vanish along points of {α
0
}.
We thus choose α
0
= ξ
0
+ i δ, δ > 0, so that (VIII.4.13) furnishes
u
j
(x, t) =
e
δ t
(2π)
2
Z
R
4
e
ix·ξ+it α
0
(δ
ij
ξ
2
−ξ
j
ξ
k
) G
j
(ξ, α
0
−i δ)
(i α
0
+ δ + ξ
2
+ λ)ξ
2
dξdα
0
.
(VIII.4.14)
Moreover, we obtain
G(ξ, α
0
− i δ) =
1
(2π)
2
Z
R
4
F (x, t)e
−δ t
e
−ix·ξ−it α
0
dx dt ,
which, by Parseval’s theorem, furnishes
Z
R
4
|G(ξ, α
0
−i δ)|
2
dξ dα
0
=
Z
T
0
dt
Z
R
3
|F (x, t)|
2
e
−δt
dx ,
where we have also employed the fact that F is of compact support in (0, T ).
Using this information in (VIII.4.14) along with the Schwarz i nequality, we
deduce
|u(x, t)| ≤ c
1
e
δ t
Z
R
4
dζ dξ
ζ
2
+ ξ
4
+ δ
2
+ λ
2
1
2
Z
T
0
Z
R
3
|F (x, t)|
2
e
−δt
!
1
2
≤ c
2
e
δ t
,
with c
2
independent of δ. If we take t < 0 in this latter relation and let δ → ∞,
we conclude that u(x, t) = 0 for all t < 0 and x ∈ R
3
, and so, by the continuity