Some Remarks on the Erdos Distinct Subset Sums Problem
Abstract
Let \a1, …, an\ ⊂ N be a set of positive integers, an denoting the largest element, so that for any two of the 2n subsets the sum of all elements is distinct. Erdos asked whether this implies an ≥ c · 2n for some universal c>0. We prove, slightly extending a result of Elkies, that for any a1, …, an ∈ R>0 ∫R ( x x )2 Πi=1n ( ai x)2 dx ≥ π2n with equality if and only if all subset sums are 1-separated. This leads to a new proof of the currently best lower bound an ≥ 2/π n · 2n. The main new insight is that having distinct subset sums and an small requires the random variable X = a1 a2 … an to be close to Gaussian in a precise sense.
Turn this paper into a lesson
ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.