uniolympiads English

IMC 2004, Ημέρα 1, Θέμα 6

For every complex number z∉{0,1}z \notin \{0, 1\} define

f(z):=∑(log⁡z)−4,f(z) := \sum (\log z)^{-4},

where the sum is over all branches of the complex logarithm.

a) Show that there are two polynomials PP and QQ such that f(z)=P(z)/Q(z)f(z) = P(z)/Q(z) for all z∈C∖{0,1}z \in \mathbb{C} \setminus \{0, 1\}. (10 points)

b) Show that for all z∈C∖{0,1}z \in \mathbb{C} \setminus \{0, 1\}

f(z)=zz2+4z+16(z−1)4.(10 points)f(z) = z\frac{z^2 + 4z + 1}{6(z-1)^4}. \qquad \text{(10 points)}
Πρωτότυπο από το επίσημο PDFFor every complex number z ∈ / {0, 1} define X f (z) := (log z) −4 , where the sum is over all branches of the complex logarithm. a) Show that there are two polynomials P and Q such that f (z) = P (z)/Q(z) for all z ∈ C \ {0, 1}. points] b) Show that for all z ∈ C \ {0, 1} f (z) = z z 2 + 4z + 1 . 6(z − 1) 4 [10 [10 points]

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

Solution 1. It is clear that the left hand side is well defined and independent of the order of summation, because we have a sum of the type ∑n−4\sum n^{-4}, and the branches of the logarithms do not matter because all branches are taken. It is easy to check that the convergence is locally uniform on C∖{0,1}\mathbb{C} \setminus \{0,1\}; therefore, ff is a holomorphic function on the complex plane, except possibly for isolated singularities at 0 and 1. (We omit the detailed estimates here.)

The function log has its only (simple) zero at z=1z = 1, so ff has a quadruple pole at z=1z = 1.

Now we investigate the behavior near infinity. We have Re(log⁡(z))=log⁡∣z∣\mathrm{Re}(\log(z)) = \log|z|, hence (with c:=log⁡∣z∣c := \log|z|)

∣∑(log⁡z)−4∣≤∑∣log⁡z∣−4=∑(log⁡∣z∣+2πin)−4+O(1)=∫−∞∞(c+2πix)−4 dx+O(1)=c−4∫−∞∞(1+2πix/c)−4 dx+O(1)=c−3∫−∞∞(1+2πit)−4 dt+O(1)≤α(log⁡∣z∣)−3\begin{aligned} |\sum(\log z)^{-4}| &\le \sum|\log z|^{-4} = \sum(\log|z| + 2\pi in)^{-4} + O(1) \\ &= \int_{-\infty}^{\infty}(c + 2\pi ix)^{-4}\,dx + O(1) \\ &= c^{-4}\int_{-\infty}^{\infty}(1 + 2\pi ix/c)^{-4}\,dx + O(1) \\ &= c^{-3}\int_{-\infty}^{\infty}(1 + 2\pi it)^{-4}\,dt + O(1) \\ &\le \alpha(\log|z|)^{-3} \end{aligned}
for a universal constant α\alpha. Therefore, the infinite sum tends to 0 as ∣z∣→∞|z| \to \infty. In particular, the isolated singularity at ∞\infty is not essential, but rather has (at least a single) zero at ∞\infty.

The remaining singularity is at z=0z = 0. It is readily verified that f(1/z)=f(z)f(1/z) = f(z) (because log⁡(1/z)=−log⁡(z)\log(1/z) = -\log(z)); this implies that ff has a zero at z=0z = 0.

We conclude that the infinite sum is holomorphic on C\mathbb{C} with at most one pole and without an essential singularity at ∞\infty, so it is a rational function, i.e. we can write f(z)=P(z)/Q(z)f(z) = P(z)/Q(z) for some polynomials PP and QQ which we may as well assume coprime. This solves the first part.

Since ff has a quadruple pole at z=1z = 1 and no other poles, we have Q(z)=(z−1)4Q(z) = (z-1)^4 up to a constant factor which we can as well set equal to 1, and this determines PP uniquely. Since f(z)→0f(z) \to 0 as z→∞z \to \infty, the degree of PP is at most 3, and since P(0)=0P(0) = 0, it follows that P(z)=z(az2+bz+c)P(z) = z(az^2 + bz + c) for yet undetermined complex constants a,b,ca, b, c.

There are a number of ways to compute the coefficients a,b,ca, b, c, which turn out to be a=c=1/6a = c = 1/6, b=2/3b = 2/3. Since f(z)=f(1/z)f(z) = f(1/z), it follows easily that a=ca = c. Moreover, the fact lim⁡z→1(z−1)4f(z)=1\lim\limits_{z\to1}(z-1)^4 f(z) = 1 implies a+b+c=1a + b + c = 1 (this fact follows from the observation that at z=1z = 1, all summands cancel pairwise, except the principal branch which contributes a quadruple pole). Finally, we can calculate

f(−1)=π−4∑n oddn−4=2π−4∑n≥1, n oddn−4=2π−4(∑n≥1n−4−∑n≥1, n evenn−4)=148 .f(-1) = \pi^{-4}\sum_{n\text{ odd}} n^{-4} = 2\pi^{-4}\sum_{n\ge1,\ n\text{ odd}} n^{-4} = 2\pi^{-4}\left(\sum_{n\ge1} n^{-4} - \sum_{n\ge1,\ n\text{ even}} n^{-4}\right) = \frac{1}{48}\,.
This implies a−b+c=−1/3a - b + c = -1/3. These three equations easily yield a,b,ca, b, c.

Moreover, the function ff satisfies f(z)+f(−z)=16f(z2)f(z) + f(-z) = 16f(z^2): this follows because the branches of log⁡(z2)=log⁡((−z)2)\log(z^2) = \log((-z)^2) are the numbers 2log⁡(z)2\log(z) and 2log⁡(−z)2\log(-z). This observation supplies the two equations b=4ab = 4a and a=ca = c, which can be used instead of some of the considerations above.

Another way is to compute g(z)=∑1(log⁡z)2g(z) = \sum\frac{1}{(\log z)^2} first. In the same way, g(z)=dz(z−1)2g(z) = \frac{dz}{(z-1)^2}. The unknown coefficient dd can be computed from lim⁡z→1(z−1)2g(z)=1\lim\limits_{z\to1}(z-1)^2 g(z) = 1; it is d=1d = 1. Then the exponent 2 in the denominator can be increased by taking derivatives (see Solution 2). Similarly, one can start with exponent 3 directly.

A more straightforward, though tedious way to find the constants is computing the first four terms of the Laurent series of ff around z=1z = 1. For that branch of the logarithm which vanishes at 1, for all ∣w∣<12|w| < \frac{1}{2} we have

log⁡(1+w)=w−w22+w33−w44+O(∣w∣5);\log(1+w) = w - \frac{w^2}{2} + \frac{w^3}{3} - \frac{w^4}{4} + O(|w|^5);
after some computation, one can obtain
1log⁡(1+w)4=w−4+2w−2+76w−2+16w−1+O(1).\frac{1}{\log(1+w)^4} = w^{-4} + 2w^{-2} + \frac{7}{6}w^{-2} + \frac{1}{6}w^{-1} + O(1).
The remaining branches of logarithm give a bounded function. So
f(1+w)=w−4+2w−2+76w−2+16w−1f(1+w) = w^{-4} + 2w^{-2} + \frac{7}{6}w^{-2} + \frac{1}{6}w^{-1}
(the remainder vanishes) and
f(z)=1+2(z−1)+76(z−1)2+16(z−1)3(z−1)4=z(z2+4z+1)6(z−1)4.f(z) = \frac{1 + 2(z-1) + \frac{7}{6}(z-1)^2 + \frac{1}{6}(z-1)^3}{(z-1)^4} = \frac{z(z^2 + 4z + 1)}{6(z-1)^4}.

Solution 2. ¿From the well-known series for the cotangent function,

lim⁡N→∞∑k=−NN1w+2πi⋅k=i2cot⁡iw2\lim_{N\to\infty}\sum_{k=-N}^{N}\frac{1}{w + 2\pi i\cdot k} = \frac{i}{2}\cot\frac{iw}{2}
and
lim⁡N→∞∑k=−NN1log⁡z+2πi⋅k=i2cot⁡ilog⁡z2=i2⋅ie2i⋅ilog⁡z2+1e2i⋅ilog⁡z2−1=12+1z−1.\lim_{N\to\infty}\sum_{k=-N}^{N}\frac{1}{\log z + 2\pi i\cdot k} = \frac{i}{2}\cot\frac{i\log z}{2} = \frac{i}{2}\cdot i\frac{e^{2i\cdot\frac{i\log z}{2}} + 1}{e^{2i\cdot\frac{i\log z}{2}} - 1} = \frac{1}{2} + \frac{1}{z-1}.
Taking derivatives we obtain
∑1(log⁡z)2=−z⋅(12+1z−1)′=z(z−1)2,∑1(log⁡z)3=−z2⋅(z(z−1)2)′=z(z+1)2(z−1)3\begin{aligned} \sum\frac{1}{(\log z)^2} &= -z\cdot\left(\frac{1}{2} + \frac{1}{z-1}\right)' = \frac{z}{(z-1)^2}, \\ \sum\frac{1}{(\log z)^3} &= -\frac{z}{2}\cdot\left(\frac{z}{(z-1)^2}\right)' = \frac{z(z+1)}{2(z-1)^3} \end{aligned}
and
∑1(log⁡z)4=−z3⋅(z(z+1)2(z−1)3)′=z(z2+4z+1)2(z−1)4.\sum\frac{1}{(\log z)^4} = -\frac{z}{3}\cdot\left(\frac{z(z+1)}{2(z-1)^3}\right)' = \frac{z(z^2 + 4z + 1)}{2(z-1)^4}.

Μεταγραφή από το επίσημο κείμενο λύσεων, σελίδες 3–4. 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;