Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Claim. Let be a set of nonnegative integers such that every is with and a positive integer. Theorem 1 of Cilleruelo's paper gives
and for the constant is . If is an additive complement of the squares in the sense of Problem 33, so that every integer beyond some is with , then every integer up to any is with and in or in . Theorem 1 asks for , and a member of whose only representation is is not covered, so the theorem does not apply verbatim; adding for each such could double the count and give only . The proof applies unchanged: allowing adds to the inner sum for each one term bounded by the maximum of the continuous weight, which the error of the paper's Euler-summation lemma (Lemma 2) absorbs. Habsieger's theorem, which allows the square , gives the same bound directly. The added elements change by at most a constant, so
This answers the second question of the problem yes. The same bound was proved independently by Habsieger, whose paper notes Cilleruelo's proof in a note added in proof; the first answer, , is Moser's.
Covers. The liminf question (the part liminf), answered yes with the
bound . Not covered: the smallest possible limsup, which no source
determines; since a limsup is at least the liminf, the bound shows only that
the smallest possible limsup is at least .
Depends on. Nothing in this wiki; the claim rests on the cited paper.
Acceptance. Refereed: J. Cilleruelo, The additive completion of th
powers, J. Number Theory 44 (1993), no. 3, 237–243. The site's curator records
the bound in the problem's remarks, but the site labels the problem OPEN, so
that remark is not acceptance of the problem and the page lists no reviewed
evidence. The edition read is an author-typeset manuscript; its proof is not
compiled in this corpus.
Dating. The page is dated by the issue month in the publisher's record, July 1993; the day is a placeholder.