Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Statement. Put bn=(n+1)!+nb_n=(n+1)!+n for n∈N0n\in\mathbb N_0, and B={bn:n≥0}B=\{b_n:n\ge0\}. Then BB is Sidon and meets every infinite arithmetic progression in N0\mathbb N_0, so its complement contains none. The set BB therefore answers Problem 198 negatively.

Source. The public problem page attributes this explicit construction to AlphaProof. Its discussion links an AI-assisted formalization. The source record distinguishes that report from the natural-language proof below and from a reproduced build.

Proof. Every bnb_n is positive, and for n≥0n\ge0,

bn+1−2bn=(n+2)!+(n+1)−2((n+1)!+n)=n((n+1)!−1)+1>0.b_{n+1}-2b_n =(n+2)!+(n+1)-2((n+1)!+n) =n\bigl((n+1)!-1\bigr)+1>0.

Thus the doubling-gap lemma makes BB a Sidon set, including uniqueness of sums with equal summands.

Let P(a,d)={a+td:t≥0}P(a,d)=\{a+td:t\ge0\} with a≥0a\ge0 and d≥1d\ge1. Set n=a+dn=a+d. Since 1≤d≤n+11\le d\le n+1, the integer dd divides (n+1)!(n+1)!. Hence

bn=a+((n+1)!d+1)d∈P(a,d).b_n=a+\left(\frac{(n+1)!}{d}+1\right)d\in P(a,d).

This point also belongs to BB, so P(a,d)P(a,d) cannot be contained in the complement. The case a=0a=0 is included, and all elements of BB are positive. □\square

Method. The index fixes a residue class while factorial divisibility removes the nonlinear term modulo any prescribed step. Rapid growth prevents collisions of two-term sums. This is an explicit instance of the same separation mechanism as the enumeration construction.

Bears on. Problem 198: the set BB is a Sidon set whose complement contains no infinite arithmetic progression, a negative answer.