Bourbaki

Theorem 4

Tag 08C5, permanent. Cite it as [Bourbaki, Tag 08C5].

Functions of a Real Variable, chapter IV, § 1. EXISTENCE THEOREMS

Let ff be a locally Lipschitz function on I×HI \times H and (a,b)(a, b) an arbitrary point of I×HI \times H.

1 There exist an interval K⊂IK \subset I, a neighbourhood of aa in II, and a neighbourhood VV of bb in HH such that, for every point (t0,x0)(t_0, x_0) of K×VK \times V, there exists an integral (and only one) u(t,t0,x0)u(t, t_0, x_0) defined on KK, with values in HH and equal to x0x_0 at the point t0t_0 (in other words, J(t0,x0)⊃KJ(t_0, x_0) \supset K for all (t0,x0)∈K×V(t_0, x_0) \in K \times V).

2 The map (t,t0,x0)↦u(t,t0,x0)(t, t_0, x_0) \mapsto u(t, t_0, x_0) of K×K×VK \times K \times V into HH is uniformly continuous.

3 There exists a neighbourhood W⊂VW \subset V of bb in HH such that, for every point

(t,t0,x0)∈K×K×W, (t, t_0, x_0) \in K \times K \times W,
the equation x0=u(t0,t,x)x_0 = u(t_0, t, x) has a unique solution xx on VV equal to u(t,t0,x0)u(t, t_0, x_0) ("resolution of the integral with respect to the constant of integration").

1 Let SS be a ball with centre bb and radius rr contained in HH, and J0J_0 an interval contained in II, a neighbourhood of aa in II, such that ff is bounded and Lipschitz (with respect to some constant kk) on J0×SJ_0 \times S; denote by MM the supremum of ∥f(t,x)∥\|f(t, x)\| on J0×SJ_0 \times S. Then there exist (IV, p. 171, th. 1) an interval J⊂J0J \subset J_0, a neighbourhood of aa in II, and an integral vv of x′=f(t,x)x' = f(t, x) defined on JJ, with values in SS and equal to bb at aa. We shall see that the open ball VV with centre bb and radius r/2r/2, and the intersection KK of JJ with an interval ]a−l,a+l[]a - l, a + l[, where ll is small enough, are as required. Indeed, prop. 7 of IV, p. 174 (applied to the set J0×SJ_0 \times S and the case where α=0\alpha = 0) shows that there exists an integral of x′=f(t,x)x' = f(t, x) defined on KK, with values in SS, and equal to x0x_0 at a point t0∈Kt_0 \in K, provided that

∥v(t)−b∥+∥v(t0)−x0∥ek∣t−t0∣<r \|v(t) - b\| + \|v(t_0) - x_0\| e^{k|t-t_0|} < r
for every t∈Kt \in K. Now, by the mean value theorem, one has
∥v(t)−b∥≤M∣t−a∣≤Ml \|v(t) - b\| \leq M |t - a| \leq Ml
for every t∈Kt \in K; since ∥x0−b∥<r/2\|x_0 - b\| < r/2 one sees that it suffices to take ll such that
Ml+(Ml+r/2)e2kl<r Ml + (Ml + r/2)e^{2kl} < r
or the relation (18) to be satisfied for every (t,t0,x0)(t, t_0, x_0) of K×K×VK \times K \times V.

2) By the mean value theorem we have

∥u(t1,t0,x0)−u(t2,t0,x0)∥≤M∣t2−t1∣ \| \mathbf{u}(t_1, t_0, \mathbf{x}_0) - \mathbf{u}(t_2, t_0, \mathbf{x}_0) \| \leq M |t_2 - t_1|

(20)

for all t0,t1,t2t_0, t_1, t_2 in K and x0\mathbf{x}_0 in V. Now prop. 5 (IV, p. 169) shows that

∥u(t,t0,x1)−u(t,t0,x2)∥≤e2kt ∥x2−x1∥ \| \mathbf{u}(t, t_0, \mathbf{x}_1) - \mathbf{u}(t, t_0, \mathbf{x}_2) \| \leq e^{2kt} \ \| \mathbf{x}_2 - \mathbf{x}_1 \|

(21)

for every tt and t0t_0 in K, and x1\mathbf{x}_1 and x2\mathbf{x}_2 in V. Finally, if t1t_1 and t2t_2 are any two points in K,

∥u(t1,t2,x0)−u(t2,t2,x0)∥≤M∣t2−t1∣ \| \mathbf{u}(t_1, t_2, \mathbf{x}_0) - \mathbf{u}(t_2, t_2, \mathbf{x}_0) \| \leq M |t_2 - t_1|

by the mean value theorem, that is to say

∥u(t1,t2,x0)−x0∥≤M∣t2−t1∣; \| \mathbf{u}(t_1, t_2, \mathbf{x}_0) - \mathbf{x}_0 \| \leq M |t_2 - t_1| ;

since u(t,t2,x0)\mathbf{u}(t, t_2, \mathbf{x}_0) is identical to the integral which takes the value u(t1,t2,x0)\mathbf{u}(t_1, t_2, \mathbf{x}_0) at the point t1t_1, prop. 5 (IV, p. 169) shows that

∥u(t,t1,x0)−u(t,t2,x0)∥≤Me2kl∣t2−t1∣ \| \mathbf{u}(t, t_1, \mathbf{x}_0) - \mathbf{u}(t, t_2, \mathbf{x}_0) \| \leq Me^{2kl} |t_2 - t_1|

(22)

for all t,t1,t2t, t_1, t_2 in K and x0\mathbf{x}_0 in V. The three inequalities (20), (21) and (22) thus prove the uniform continuity of the map (t,t0,x0)↦u(t,t0,x0)(t, t_0, \mathbf{x}_0) \mapsto \mathbf{u}(t, t_0, \mathbf{x}_0) on K×K×VK \times K \times V.

3 By (20), we have ∥u(t,t0,x0)−x0∥≤M∣t−t0∣≤2Ml\| \mathbf{u}(t, t_0, \mathbf{x}_0) - \mathbf{x}_0 \| \leq M |t - t_0| \leq 2Ml on

K×K×V. K \times K \times V.

If ll is taken small enough, so that 2Ml<r/42Ml < r/4, one then sees that if x0\mathbf{x}_0 is any point of the open ball W with centre b\mathbf{b} and radius r/4r/4, that u(t,t0,x0)∈V\mathbf{u}(t, t_0, \mathbf{x}_0) \in V for any tt and t0t_0 in K. If x=u(t,t0,x0)\mathbf{x} = \mathbf{u}(t, t_0, \mathbf{x}_0), the function s↦u(s,t,x)s \mapsto \mathbf{u}(s, t, \mathbf{x}) is then defined on K and is equal to the integral of (1) which takes the value x\mathbf{x} at the point tt, that is, to u(s,t0,x0)\mathbf{u}(s, t_0, \mathbf{x}_0); in particular

x0=u(t0,t0,x0)=u(t0,t,x). \mathbf{x}_0 = \mathbf{u}(t_0, t_0, \mathbf{x}_0) = \mathbf{u}(t_0, t, \mathbf{x}).

Moreover, if y∈Vy \in V is such that x0=u(t0,t,y)\mathbf{x}_0 = \mathbf{u}(t_0, t, y), then the integral s↦u(s,t,y)s \mapsto \mathbf{u}(s, t, y) takes the value x0\mathbf{x}_0 at t0t_0 so is identical to s↦u(s,t0,x0)s \mapsto \mathbf{u}(s, t_0, \mathbf{x}_0), which consequently takes the value x\mathbf{x} at tt, which shows that y=xy = \mathbf{x} and concludes the proof.

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Transcribed from the printed volume. Not translated.