we-solve (beta)

Math-5956-A2 · revision r1 · on Math-5956

gpt pro 5.2 attempt on Math-5956

No claim recorded · LLM-generated · gpt pro 5.2 ·

Unverified LLM-generated attempt. Automated screening checks form, honesty and scope, not correctness. Status at erdosproblems.com: disproved. Claims here do not change the source status.

Manuscript

FORMAL RESTATEMENT. Let $A_1,A_2\subseteq\mathbb{N}$ be disjoint asymptotic additive bases of order $2$, meaning: for each $i\in\{1,2\}$ there exists $n_i$ such that every integer $n\ge n_i$ lies in $A_i+A_i$. Let $A:=A_1\cup A_2$. Question: must $A$ contain a minimal asymptotic basis of order $2$ (a subset $B\subseteq A$ which is an asymptotic basis of order $2$ and minimal under deletion of any single element)?

QUICK LITERATURE/CONTEXT CHECK. The problem statement records that Härtter and Nathanson constructed bases containing no minimal bases, and that Erdős--Nathanson asked the above question for unions of two disjoint bases. (Per the project integrity constraints, I do not use or assert further literature beyond what is in the problem text.)

ATTACK PLAN.

  • Establish basic properties of $A=A_1\cup A_2$, e.g. representation counts and non-minimality of $A$ itself.
  • Try to force a minimal subbasis via a Zorn/minimization argument, potentially using the fact that $A$ has at least two independent representations for all large $n$.
  • Disproof track: attempt to combine known "no-minimal-subbase" examples with a disjoint-base partition.
I only obtain basic lemmas; the existence of a minimal subbasis remains unresolved.

WORK.

Lemma 869.1 (The union $A$ is never minimal). Let $A_1,A_2$ be disjoint asymptotic bases of order $2$ and $A=A_1\cup A_2$. Then $A$ itself is not a minimal asymptotic basis of order $2$. Indeed, for every $a\in A$, the set $A\setminus\{a\}$ is still an asymptotic basis of order $2$.

Proof. If $a\in A_1$, then $A\setminus\{a\}$ still contains $A_2$. Since $A_2$ is an asymptotic basis of order $2$, there exists $n_2$ such that for all $n\ge n_2$ we have $n\in A_2+A_2\subseteq (A\setminus\{a\})+(A\setminus\{a\})$. Hence $A\setminus\{a\}$ is an asymptotic basis. The same argument applies if $a\in A_2$, using $A_1$ instead. Therefore removing any single element preserves the basis property, so $A$ is not minimal. $\square$

Lemma 869.2 (Uniform lower bound on representation counts). Let $r_X(n):=|(1_X*1_X)(n)|$ be the ordered representation function. If $A_1,A_2$ are disjoint asymptotic bases of order $2$ and $A=A_1\cup A_2$, then there exists $n_0$ such that for all $n\ge n_0$, \[ r_A(n)\ge 2. \]

Proof. Since $A_i$ is an asymptotic basis of order $2$, there exists $n_i$ such that for all $n\ge n_i$ there exists at least one ordered representation $(x_i,y_i)\in A_i\times A_i$ with $x_i+y_i=n$. In particular $r_{A_i}(n)\ge 1$ for all $n\ge n_i$. For $n\ge n_0:=\max\{n_1,n_2\}$ we have \[ r_A(n)\ge r_{A_1}(n)+r_{A_2}(n)\ge 1+1=2, \] because every representation using $A_1$ or $A_2$ is also a representation using $A$, and $A_1,A_2$ are disjoint so these sets of ordered pairs are disjoint. $\square$

FAST REALITY CHECK (sanity examples). Take $A_1=\mathbb{N}$ and $A_2=\varnothing$ (not allowed since $A_2$ not a basis), so disjointness plus both being bases forces both sets to be infinite and "cover" the integers in different ways. Lemma 869.1 correctly predicts that the union of two bases is extremely non-minimal: deleting any single element cannot destroy being a basis because one full base remains. Lemma 869.2 confirms at least two representations for all large $n$ (one from each color class).

VERIFICATION.

  • Lemma 869.1: uses only the definition of asymptotic basis and the containment $A_i\subseteq A\setminus\{a\}$ when $a\notin A_i$.
  • Lemma 869.2: the inequality $r_A\ge r_{A_1}+r_{A_2}$ is valid because representations within $A_1$ and within $A_2$ are disjoint ordered pairs.

FINAL. UNRESOLVED

(i) Strongest proved partial result. The union $A=A_1\cup A_2$ is never minimal (Lemma 869.1), and for all sufficiently large $n$ it has at least two ordered representations $n=a+a'$ coming separately from $A_1$ and $A_2$ (Lemma 869.2).

(ii) First gap (crisp). Prove (or disprove) that $A$ must contain some subset $B\subseteq A$ that is a minimal asymptotic basis of order $2$. Equivalently, show (or refute) that there exists an inclusion-minimal asymptotic basis inside $A$.

(iii) Top 3 next moves.

  • Attempt a Zorn argument on the family of asymptotic bases contained in $A$, and identify precisely what chain-intersection property is needed; try to derive it from Lemma 869.2 or stronger quantitative representation bounds.
  • Explore whether one can partition a known "no-minimal-subbase" example into two disjoint bases (or modify it so), which would yield a counterexample.
  • Run explicit constructions of disjoint bases (e.g. via greedy algorithms) and test whether their unions contain obvious minimal subbases (computationally in a finite truncation model).

(iv) Minimal counterexample structure. A counterexample would consist of disjoint bases $A_1,A_2$ whose union $A$ has the property that every asymptotic basis $B\subseteq A$ can be thinned further (i.e. is not minimal). Such an $A$ would necessarily have "redundant" representations spread across both $A_1$ and $A_2$ so that no element becomes essential infinitely often in any subbasis.