Documentation

FormalConjectures.ErdosProblems.«168»

Erdős Problem 168 #

Reference: erdosproblems.com/168

Say a finite set of natural numbers is non ternary if it contains no 3-term arithmetic progression of the form n, 2n, 3n.

Equations
Instances For

    IntervalNonTernarySets N is the (fin)set of non ternary subsets of {1,...,N}. The advantage of defining it as below is that some proofs (e.g. that of F 3 = 2) become rfl.

    Equations
    Instances For
      @[reducible, inline]
      abbrev Erdos168.F (N : ) :

      F N is the size of the largest non ternary subset of {1,...,N}.

      Equations
      Instances For
        theorem Erdos168.F_0 :
        F 0 = 0
        theorem Erdos168.F_1 :
        F 1 = 1
        theorem Erdos168.F_2 :
        F 2 = 2
        theorem Erdos168.F_3 :
        F 3 = 2

        Sanity check: elements of IntervalNonTernarySets N are precisely non ternary subsets of {1,...,N}

        theorem Erdos168.F_eq_card (N : ) (S : Finset ) (hS : S Finset.Icc 1 N) (hS' : NonTernary S) (hS'' : TFinset.Icc 1 N, NonTernary TS.card T.cardT.card = S.card) :
        F N = S.card

        Sanity check: if S is a maximal non ternary subset of {1,..., N} then F N is given by the cardinality of S

        theorem Erdos168.erdos_168.parts.i :
        Filter.Tendsto (fun (N : ) => (F N) / N) Filter.atTop (nhds sorry)

        What is the limit $F(N)/N$ as $N \to \infty$?

        theorem Erdos168.erdos_168.parts.ii :
        Irrational (Filter.limsup (fun (N : ) => (F N) / N) Filter.atTop) sorry

        Is the limit $F(N)/N$ as $N \to \infty$ irrational?

        theorem Erdos168.erdos_168.variants.limit_exists :
        ∃ (x : ), Filter.Tendsto (fun (N : ) => (F N) / N) Filter.atTop (nhds x)

        The limit $F(N)/N$ as $N \to \infty$ exists. (proved by Graham, Spencer, and Witsenhausen)