Positive limiting proportions for prescribed missing sums and differences

About 20 years old · traced to

For n≥1n\ge 1, let SS range over subsets of {0,1,…,n−1}\{0,1,\dots,n-1\}. For nonnegative integers jj and kk, define

ρj,k:=lim⁡n→∞2−n#{S⊂{0,1,…,n−1}:∣S+S∣=2n−1−j, ∣S−S∣=2n−1−k},\rho_{j,k}:=\lim_{n\to\infty}2^{-n}\#\{S\subset\{0,1,\dots,n-1\}:|S+S|=2n-1-j,\ |S-S|=2n-1-k\},

assuming the limit exists. Since S−SS-S is symmetric about 00, only even kk can occur. The limiting-distribution conjecture. For every pair of nonnegative integers jj and kk with kk even, ρj,k\rho_{j,k} exists and is positive; moreover,

∑j=0∞∑k=0\k even∞ρj,k=1.\sum_{j=0}^{\infty}\sum_{\substack{k=0\k\text{ even}}}^{\infty}\rho_{j,k}=1.

This refines the conjecture on the three aggregate limiting proportions by recording the exact numbers of missing sums and differences. The normalization asserts that these limiting frequencies account for all subsets asymptotically; the source also notes that the claim is known when j≥k/2j\ge k/2.

References

Primary source

Greg Martin and Kevin O'Bryant, “Many sets have more sums than differences”, arXiv:math/0608131 (2006).

Progress summary

Never refreshed

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

Solutions 0

No solutions have been posted yet.