3
$\begingroup$

When studying the $\text{UMD}$ constants of spaces like $L_{p_1}(L_{p_2}(\cdots (L_{p_n})\cdots))$, I encounter the following question: Let $\alpha > 0$, define $$C(\alpha) : = \sup_{a > 0, b> 0, 0< \delta<1} \frac{(1 + a)^{1+ \alpha} + (1 + b)^{1 + \alpha}}{(1 + \delta^2 + 2 \delta + a)^{1+\alpha} + (1 + \delta^2 - 2 \delta + b)^{1 + \alpha}}.$$

My QUESTION is:

Does there exist $M> 0$ such that for $\alpha$ sufficiently small, we have $$C(\alpha) \ge 1 + M \alpha?$$ By taking $a = 0$ in the supremum, I can show that there exists $m > 1$ such that $$C(\alpha) \ge (1 + m^{-\frac{1}{\alpha}})^\alpha,$$ this of course does not answer the preceding question.

$\endgroup$
2
  • $\begingroup$ Is this question somehow connected to a lower estimate of $\|P\|$, where P is the projection from your JFA paper on a non-UMD space? $\endgroup$
    – user42438
    Commented Nov 6, 2013 at 21:14
  • $\begingroup$ @ Mark Veraar Yes, it is connected, it is some quantity a priori larger (or maybe not strictly larger) than the norm $\| P\|$ in that paper. I also tried to find the explicit formula for $\| P\|$, which will probably be enough. $\endgroup$
    – Yanqi QIU
    Commented Nov 7, 2013 at 22:09

0

You must log in to answer this question.

Browse other questions tagged .