Erdős Problem #47 — Unit-fraction representations of 1 in dense sequences

Erdős

Perhaps this is really a Turán type problem and not a Ramsey problem. In other words, if mm is sufficiently large and 1<a1<a2<<am1 < a_1 < a_2 < \cdots < a_\ell \leq m is a sequence of integers for which 1/a>δlogm\sum_\ell 1/a_\ell > \delta \log m then

1=εiai(εi=0 or 1)1 = \sum \frac{\varepsilon_i}{a_i} \qquad (\varepsilon_i = 0 \text{ or } 1)

is always solvable. I offer 100 dollars for a proof or disproof. Perhaps it suffices to assume that

ai<m1ai>C(loglogm)2\sum_{a_i < m} \frac{1}{a_i} > C(\log\log m)^2

for some large enough CC.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Solutions 0

No solutions have been posted yet.