Bourbaki

Lemma 1

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

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

Let ff be a map from I×HI \times H into EE such that, on writing fxf_x (for every x∈Hx \in H) for the map t↦f(t,x)t \mapsto f(t, x) of II into EE, the following conditions are satisfied: 1° fxf_x is regulated on II for every x∈Hx \in H; 2° the map x↦fxx \mapsto f_x of HH into the set F(I,E)\mathscr{F}(I, E) of maps from II into EE is continuous when one endows F(I,E)\mathscr{F}(I, E) with the topology of compact convergence (Gen. Top., X, p. 278). Under these conditions:

1° For every continuous map uu of II into HH the function t↦f(t,u(t))t \mapsto f(t, u(t)) is regulated on II; more precisely, the right (resp. left) limit of this function at a point t0∈It_0 \in I is equal to the right (resp. left) limit of the function t↦f(t,u(t0))t \mapsto f(t, u(t_0)) at the point t0t_0.

2° If (un)(u_n) is a sequence of maps of II into HH which converges uniformly to a continuous function uu of II into HH on every compact subset of II, then the sequence of functions t↦f(t,un(t))t \mapsto f(t, u_n(t)) converges uniformly to f(t,u(t))f(t, u(t)) on every compact subset of II.

1 Let cc be the right limit of f(t,u(t0))f(t, u(t_0)) at the point t0t_0; for every ε>0\varepsilon > 0 there is a compact neighbourhood VV of t0t_0 in II such that ∥f(t,u(t0))−c∥≤ε\|f(t, u(t_0)) - c\| \leq \varepsilon for t∈Vt \in V and t>t0t > t_0. On the other hand, there exists δ>0\delta > 0 such that the relations

x∈H,∥x−u(t0)∥≤δ x \in H, \quad \|x - u(t_0)\| \leq \delta

imply ∥f(s,x)−f(s,u(t0))∥≤ε\|f(s, x) - f(s, u(t_0))\| \leq \varepsilon for all s∈Vs \in V; if W⊂VW \subset V is a neighbourhood of t0t_0 in II such that ∥u(t)−u(t0)∥≤δ\|u(t) - u(t_0)\| \leq \delta for every t∈Wt \in W, then ∥f(t,u(t))−c∥≤2ε\|f(t, u(t)) - c\| \leq 2\varepsilon for t∈Wt \in W and t>t0t > t_0, which proves that cc is the right limit of f(t,u(t))f(t, u(t)) at the point t0t_0.

2 Let KK be a compact subset of II; since uu is continuous on II, u(K)u(K) is a compact subset of HH; for every ε>0\varepsilon > 0 and x∈u(K)x \in u(K) there exists a number δx\delta_x such that, for every y∈H,∥y−x∥≤δxy \in H, \|y - x\| \leq \delta_x and every t∈Kt \in K, one has ∥f(t,y)−f(t,x)∥≤ε\|f(t, y) - f(t, x)\| \leq \varepsilon. There is a finite number of points xi∈u(K)x_i \in u(K) such that the closed balls with centre xix_i and radius 12δxi\frac{1}{2}\delta_{x_i} form a cover of u(K)u(K); let δ=Min⁡(δxi)\delta = \operatorname{Min}(\delta_{x_i}). By hypothesis there exists an integer n0n_0 such that for n≥n0n \geq n_0 one has ∥un(t)−u(t)∥≤12δ\|u_n(t) - u(t)\| \leq \frac{1}{2}\delta for every t∈Kt \in K. Now, for every t∈Kt \in K there exists an index ii such that

∥u(t)−xi∥≤12δxi; \|u(t) - x_i\| \leq \frac{1}{2}\delta_{x_i};

consequently one has ∥un(t)−xi∥≤δx\|u_n(t) - x_i\| \leq \delta_x, whence ∥f(t,un(t))−f(t,u(t))∥≤2ε\|f(t, u_n(t)) - f(t, u(t))\| \leq 2\varepsilon for every t∈Kt \in K and every n≥n0n \geq n_0

For the rest of this section I will denote an interval contained in R\mathbf{R}, not reducing to a single point, HH an open nonempty set contained in the normed space EE, and ff a map from I×HI \times H into EE satisfying the hypotheses of lemma 1.

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

Cited by

Nothing.

Cites

Nothing.

Transcribed from the printed volume. Not translated.