uniolympiads Ελληνικά

IMC 2017, Day 2, Problem 10

Let KK be an equilateral triangle in the plane. Prove that for every p>0p > 0 there exists an ε>0\varepsilon > 0 with the following property: If nn is a positive integer, and T1,…,TnT_1, \dots, T_n are non-overlapping triangles inside KK such that each of them is homothetic to KK with a negative ratio, and

∑ℓ=1narea⁡(Tℓ)>area⁡(K)−ε,\sum_{\ell=1}^{n} \operatorname{area}(T_\ell) > \operatorname{area}(K) - \varepsilon,

then

∑ℓ=1nperimeter⁡(Tℓ)>p.\sum_{\ell=1}^{n} \operatorname{perimeter}(T_\ell) > p.
Original from the official PDFLet K be an equilateral triangle in the plane. Prove that for every p > 0 there exists an ε > 0 with the following property: If n is a positive integer, and T 1 , . . . , T n are non-overlapping triangles inside K such that each of them is homothetic to K with a negative ratio, and Problem 10. n X area (T ` ) > area (K) − ε, `=1 then n X perimeter (T ` ) > p. `=1

This transcription was compared symbol by symbol with the original, but not yet by a person.

Time on this problem 0:00

Work for at least 45 minutes before opening the solution.

Official solution

(Proposed by Fedor Malyshev, Steklov Math. Inst. and Ilya Bogdanov, MIPT, Moscow)

Solution. For an arbitrary ε>0\varepsilon > 0 we will establish a lower bound for the sum of perimeters that would tend to +∞+\infty as ε→+0\varepsilon\to +0; this solves the problem.

Rotate and scale the picture so that one of the sides of KK is the segment from (0,0)(0,0) to (0,1)(0,1), and stretch the picture horizontally in such a way that the projection of KK to the xx axis is [0,1][0,1]. Evidently, we may work with the lengths of the projections to the xx or yy axis instead of the perimeters and consider their sum, that is why we may make any affine transformation.

Let fi(a)f_i(a) be the length of intersection of the straight line {x=a}\{x = a\} with TiT_i and put f(a)=∑ifi(a)f(a) = \sum_i f_i(a). Then ff is piece-wise increasing with possible downward gaps, f(a)≤1−af(a)\le 1-a, and

∫01f(x) dx≥12−ε.\int_0^1 f(x)\,\mathrm{d}x \ge \frac{1}{2} - \varepsilon.

Let d1,…,dNd_1,\ldots,d_N be the values of the gaps of ff. Every gap is a sum of side-lengths of some of TiT_i and every TiT_i contributes to one of djd_j, we therefore estimate the sum of the gaps of ff.

In the points of differentiability of ff we have f′(a)≥f(a)/af'(a)\ge f(a)/a; this follows from fi′(a)≥fi(a)/af_i'(a)\ge f_i(a)/a after summation. Indeed, if fif_i is zero this inequality holds trivially, and if not then fi′=1f_i' = 1 and the inequality reads fi(a)≤af_i(a)\le a, which is clear from the definition.

Choose an integer m=⌊1/(8ε)⌋m = \lfloor 1/(8\varepsilon)\rfloor (considering ε\varepsilon sufficiently small). Then for all k=0,1,…,[(m−1)/2]k = 0,1,\ldots,[(m-1)/2] in the section of KK by the strip k/m≤x≤(k+1)/mk/m\le x\le (k+1)/m the area, covered by the small triangles TiT_i is no smaller than 1/(2m)−ε≥1/(4m)1/(2m) - \varepsilon\ge 1/(4m). Thus

∫k/m(k+1)/mf′(x) dx≥∫k/m(k+1)/mf(x) dxx≥mk+1∫k/m(k+1)/mf(x) dx≥mk+1⋅14m=14(k+1).\int_{k/m}^{(k+1)/m} f'(x)\,\mathrm{d}x \ge \int_{k/m}^{(k+1)/m}\frac{f(x)\,\mathrm{d}x}{x} \ge \frac{m}{k+1}\int_{k/m}^{(k+1)/m} f(x)\,\mathrm{d}x \ge \frac{m}{k+1}\cdot\frac{1}{4m} = \frac{1}{4(k+1)}.
Hence,
∫01/2f′(x) dx≥14(11+⋯+1[(m−1)/2]).\int_0^{1/2} f'(x)\,\mathrm{d}x \ge \frac{1}{4}\left(\frac{1}{1} + \cdots + \frac{1}{[(m-1)/2]}\right).
The right hand side tends to infinity as ε→+0\varepsilon\to +0. On the other hand, the left hand side equals
f(1/2)+∑xi<1/2di;f(1/2) + \sum_{x_i<1/2} d_i;
hence ∑idi\sum_i d_i also tends to infinity.

Transcribed from the official solutions, page 5. This transcription has not been checked a second time; if in doubt, look at the original.

Archive problem: the statement is the official text (transcribed, with the original above). Subject classification and difficulty were estimated automatically; hints and a worked solution are added over time.

How did it go?