uniolympiads English

IMC 2017, Ημέρα 2, Θέμα 9

Define the sequence f1,f2,… ⁣:[0,1)→Rf_1, f_2, \dots\colon [0, 1) \to \mathbb{R} of continuously differentiable functions by the following recurrence:

f1=1;fn+1′=fnfn+1on (0,1),andfn+1(0)=1.f_1 = 1; \qquad f'_{n+1} = f_n f_{n+1} \quad \text{on } (0, 1), \quad \text{and} \quad f_{n+1}(0) = 1.

Show that lim⁡n→∞fn(x)\lim_{n \to \infty} f_n(x) exists for every x∈[0,1)x \in [0, 1) and determine the limit function.

Πρωτότυπο από το επίσημο PDFDene the sequence f 1 , f 2 , . . . : [0, 1) → R of continuously differentiable functions by the following recurrence: Problem 9. f 1 = 1; 0 f n+1 = f n f n+1 on (0, 1), and f n+1 (0) = 1. Show that lim f n (x) exists for every x ∈ [0, 1) and determine the limit function. n→∞

Η μεταγραφή συγκρίθηκε σύμβολο προς σύμβολο με το πρωτότυπο, όχι ακόμα από άνθρωπο.

Χρόνος σε αυτό το θέμα 0:00

Δούλεψε τουλάχιστον 45 λεπτά πριν ανοίξεις τη λύση.

Επίσημη λύση

(Proposed by Tomáš Bárta, Charles University, Prague)

Solution. First of all, the sequence fnf_n is well defined and it holds that

fn+1(x)=e∫0xfn(t)dt.(2)f_{n+1}(x) = e^{\int_0^x f_n(t)\mathrm{d}t}. \tag{2}

The mapping Φ:C([0,1))→C([0,1))\Phi : C([0,1))\to C([0,1)) given by

Φ(g)(x)=e∫0xg(t)dt\Phi(g)(x) = e^{\int_0^x g(t)\mathrm{d}t}
is monotone, i.e. if f<gf < g on (0,1)(0,1) then
Φ(f)(x)=e∫0xf(t)dt<e∫0xg(t)dt=Φ(g)(x)\Phi(f)(x) = e^{\int_0^x f(t)\mathrm{d}t} < e^{\int_0^x g(t)\mathrm{d}t} = \Phi(g)(x)
on (0,1)(0,1). Since f2(x)=e∫0x1mathrmdt=ex>1=f1(x)f_2(x) = e^{\int_0^x 1mathrmdt} = e^x > 1 = f_1(x) on (0,1)(0,1), we have by induction fn+1(x)>fn(x)f_{n+1}(x) > f_n(x) for all x∈(0,1)x\in(0,1), n∈Nn\in\mathbb{N}. Moreover, function f(x)=11−xf(x) = \frac{1}{1-x} is the unique solution to f′=f2f' = f^2, f(0)=1f(0) = 1, i.e. it is the unique fixed point of Φ\Phi in {φ∈C([0,1)): φ(0)=1}\{\varphi\in C([0,1)) :\ \varphi(0) = 1\}. Since f1<ff_1 < f on (0,1)(0,1), by induction we have fn+1=Φ(fn)<Φ(f)=ff_{n+1} = \Phi(f_n) < \Phi(f) = f for all n∈Nn\in\mathbb{N}. Hence, for every x∈(0,1)x\in(0,1) the sequence fn(x)f_n(x) is increasing and bounded, so a finite limit exists.

Let us denote the limit g(x)g(x). We show that g(x)=f(x)=11−xg(x) = f(x) = \frac{1}{1-x}. Obviously, g(0)=lim⁡fn(0)=1g(0) = \lim f_n(0) = 1. By f1≡1f_1\equiv 1 and (2), we have fn>0f_n > 0 on [0,1)[0,1) for each n∈Nn\in\mathbb{N}, and therefore (by (2) again) the function fn+1f_{n+1} is increasing. Since fnf_n, fn+1f_{n+1} are positive and increasing also fn+1′f'_{n+1} is increasing (due to fn+1′=fnfn+1f'_{n+1} = f_nf_{n+1}), hence fn+1f_{n+1} is convex. A pointwise limit of a sequence of convex functions is convex, since we pass to a limit n→∞n\to\infty in

fn(λx+(1−λ)y)≤λfn(x)+(1−λ)fn(y)f_n(\lambda x + (1-\lambda)y) \le \lambda f_n(x) + (1-\lambda)f_n(y)
and obtain
g(λx+(1−λ)y)≤λg(x)+(1−λ)g(y)g(\lambda x + (1-\lambda)y) \le \lambda g(x) + (1-\lambda)g(y)
for any fixed xx, y∈[0,1)y\in[0,1) and λ∈(0,1)\lambda\in(0,1). Hence, gg is convex, and therefore continuous on (0,1)(0,1). Moreover, gg is continuous in 00, since 1≡f1≤g≤f1\equiv f_1\le g\le f and lim⁡x→0+f(x)=1\lim_{x\to 0+} f(x) = 1. By Dini's Theorem, convergence fn→gf_n\to g is uniform on [0,1−ε][0,1-\varepsilon] for each ε∈(0,1)\varepsilon\in(0,1) (a monotone sequence converging to a continuous function on a compact interval). We show that Φ\Phi is continuous and therefore fnf_n have to converge to a fixed point of Φ\Phi.

In fact, let us work on the space C([0,1−ε])C([0,1-\varepsilon]) with any fixed ε∈(0,1)\varepsilon\in(0,1), ∥⋅∥\|\cdot\| being the supremum norm on [0,1−ε][0,1-\varepsilon]. Then for a fixed function hh and ∥φ−h∥<δ\|\varphi - h\| < \delta we have

sup⁡x∈[0,1−ε]∣Φ(h)(x)−Φ(φ)(x)∣=sup⁡x∈[0,1−ε]e∫0xh(t)dt∣1−e∫0xφ(t)−h(t)dt∣≤C(eδ−1)<2Cδ\sup_{x\in[0,1-\varepsilon]} |\Phi(h)(x) - \Phi(\varphi)(x)| = \sup_{x\in[0,1-\varepsilon]} e^{\int_0^x h(t)\mathrm{d}t}\left|1 - e^{\int_0^x \varphi(t)-h(t)\mathrm{d}t}\right| \le C(e^\delta - 1) < 2C\delta
for δ>0\delta > 0 small enough. Hence, Φ\Phi is continuous on C([0,1−ε])C([0,1-\varepsilon]). Let us assume for contradiction that Φ(g)≠g\Phi(g)\ne g. Hence, there exists η>0\eta > 0 and x0∈[0,1−ε]x_0\in[0,1-\varepsilon] such that ∣Φ(g)(x0)−g(x0)∣>η|\Phi(g)(x_0) - g(x_0)| > \eta. There exists δ>0\delta > 0 such that ∥Φ(φ)−Φ(g)∥<13η\|\Phi(\varphi) - \Phi(g)\| < \frac{1}{3}\eta whenever ∥φ−g∥<δ\|\varphi - g\| < \delta. Take n0n_0 so large that ∥fn−g∥<min⁡{δ,13η}\|f_n - g\| < \min\{\delta, \frac{1}{3}\eta\} for all n≥n0n\ge n_0. Hence, ∥fn+1−Φ(g)∥=∥Φ(fn)−Φ(g)∥<13η\|f_{n+1} - \Phi(g)\| = \|\Phi(f_n) - \Phi(g)\| < \frac{1}{3}\eta. On the other hand, we have ∣fn+1(x0)−Φ(g)(x0)∣>∣Φ(g)(x0)−g(x0)∣−∣g(x0)−fn+1(x0)∣>η−13η=23η|f_{n+1}(x_0) - \Phi(g)(x_0)| > |\Phi(g)(x_0) - g(x_0)| - |g(x_0) - f_{n+1}(x_0)| > \eta - \frac{1}{3}\eta = \frac{2}{3}\eta, contradiction. So, Φ(g)=g\Phi(g) = g.

Since ff is the only fixed point of Φ\Phi in {φ∈C([0,1−ε]): φ(0)=1}\{\varphi\in C([0,1-\varepsilon]) :\ \varphi(0) = 1\}, we have g=fg = f on [0,1−ε][0,1-\varepsilon]. Since ε∈(0,1)\varepsilon\in(0,1) was arbitrary, we have lim⁡n→∞fn(x)=11−x\lim_{n\to\infty} f_n(x) = \frac{1}{1-x} for all x∈[0,1)x\in[0,1).

Μεταγραφή από το επίσημο κείμενο λύσεων, σελίδες 3–5. Η μεταγραφή δεν έχει ελεγχθεί δεύτερη φορά· αν έχεις αμφιβολία, κοίτα το πρωτότυπο.

Θέμα αρχείου: η εκφώνηση είναι το επίσημο κείμενο (μεταγραμμένο, με το πρωτότυπο παραπάνω). Η ταξινόμηση ανά αντικείμενο και η δυσκολία εκτιμήθηκαν αυτόματα· υποδείξεις και λύση στα ελληνικά προστίθενται σταδιακά.

Πώς πήγε;