Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

This is a theorem of Flajolet and Odlyzko (I think): Let $f(z)$ be a function analytic in a domain $$D = \{z : |z| \leq s_1, |\text{Arg}(z-s)| > \frac{\pi}{2} - \eta \},$$ where $s, s_1 > s,$ and $\eta$ are three positive real numbers. Assume that, with $ \sigma(u) = u^\alpha \log^\beta u$ and $\alpha \notin \{0, -1, -2, \dots \}$, we have $$ f(z) \sim \sigma \left( \frac{1}{1-z/s} \right) \qquad \text{ as } z \rightarrow s \text{ in } D.$$ Then, the Taylor coefficients of $f(z)$ satisfy $$[z^n]f(z) \sim s^{-n} \frac{\sigma(n)}{n \Gamma(\alpha)}.$$

My question is whether, with the same premise, except that $\alpha = 0, \beta \in \{1, 2, 3, \dots\}$, we have $$[z^n]f(z) \sim \beta s^{-n} n^{-1} \log^{\beta-1} n.$$

share|improve this question

1 Answer 1

up vote 1 down vote accepted
+50

No. For a counterexample, take $i=\sqrt{-1}$, $$f(z)= C(1-z)^i + \log [(1-z)^{-1}],$$ and let the domain $D$ be as above with the parameters $s=1$, $s_1=2$, $\eta=\pi/4$. In $D$, we have $$|(1-z)^i| = |e^{i \log (1-z)}| = e^{-\arg (1-z)},$$ which is bounded. Therefore, as $z\to 1$ in $D$, $f(z)\sim \log [(1-z)^{-1}]$. However, by Theorem VI.2 (p. 385, [Flajolet and Sedgewick 2009]), $$[z^n]f(z)=\frac{1}{n}+C [z^n] (1-z)^i =\frac{1}{n}(1+C\frac{e^{-i\log n}}{\Gamma(-i)}+O(\frac{1}{\log n})),$$ so $[z^n]f(z)$ is not asymptotic to $1/n$.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.