The 2143 three-column Fibonacci-polynomial conjecture

From papers

Let Fs(q)F_s(q) be defined by

F0(q)=1,F1(q)=1,F_0(q)=1,\qquad F_1(q)=1,

and, for s2s\ge 2,

Fs(q)=(1+q+2q2)Fs1(q)+q3Fs2(q).F_s(q)=(1+q+2q^2)F_{s-1}(q)+q^3F_{s-2}(q).

Let ENs,t(σ)(q)EN_{s,t}(\sigma)(q) denote the inversion-generating function for pattern-avoiding linear extensions of the rectangular poset with parameters s,ts,t. The 2143 three-column Fibonacci-polynomial conjecture. For all s1s\ge 1,

ENs,3(2143)(q)=q9(s2)Fs(1q).EN_{s,3}(2143)(q)=q^{9\binom{s}{2}}F_s\left(\frac{1}{q}\right).

The claim is presented as a qq-analogue of an established formula and is supported by computations for s9s\le 9; it remains open in general.

Progress summary

Solved

The conjecture is proved: a 2019 paper establishes the proposed formula for every positive integer parameter.

The 2016 source proposed the identity ENs,3(2143)(q)=q9(s2)Fs(q1)EN_{s,3}(2143)(q)=q^{9\binom{s}{2}}F_s(q^{-1}) as Conjecture 7.3, after checking it computationally through s=9s=9.

Known results

The original paper also recorded the unweighted specialization ENs,3(2143)=F3s1\lvert EN_{s,3}(2143)\rvert=F_{3s-1} and established related formulas for other column counts.

2019 proof

Colin Defant’s Theorem 3.3 proves the conjectured identity for all s1s\geq1. The argument partitions extensions into five classes, introduces an auxiliary polynomial, and derives recurrences that normalize to the defining recurrence for Fs(q1)F_s(q^{-1}).

Current status (as of August 2026): The three-column Fibonacci-polynomial conjecture is settled for every s1s\geq1 by Defant’s published proof.

Sources
Sources & referencesView supporting material

Primary source

David Anderson, Eric S. Egge, Manda Riehl, Lucas Ryan, Ruth Steinke and Yuriko Vaughan, “Pattern Avoiding Linear Extensions of Rectangular Posets”, arXiv:1605.06825 (2016).

Solutions 1

Proof

This conjecture was proved by Colin Defant in Theorem 3.3 of “Proofs of Conjectures about Pattern-Avoiding Linear Extensions”:

https://arxiv.org/abs/1905.02309

Published version:

https://doi.org/10.23638/DMTCS-21-4-16

Here is the recurrence argument. Write

Es(q)=ENs,3(2143)(q).E_s(q)=EN_{s,3}(2143)(q).

Defant introduces Hs(q)H_s(q) as the inversion-generating polynomial for those members of ENs,3(2143)EN_{s,3}(2143) whose second entry is 3s13s-1. A partition into five forced-prefix classes, followed by deletion of the entries 3s2,3s1,3s3s-2,3s-1,3s, gives

Es=q9(s1)2(1+q+q2)Es1+q9(s1)3(1+q)Hs1,E_s = q^{9(s-1)-2}(1+q+q^2)E_{s-1} + q^{9(s-1)-3}(1+q)H_{s-1},

and

Hs=q9(s1)1(1+q)Es1+q9(s1)2Hs1,H_s = q^{9(s-1)-1}(1+q)E_{s-1} + q^{9(s-1)-2}H_{s-1},

with E1=H1=1E_1=H_1=1.

Substituting the second recurrence into the first at index s+1s+1, and then using the first recurrence at index ss to eliminate Hs1H_{s-1}, gives

Es+1=q9s2(q2+q+2)Es+q18s12Es1.E_{s+1} = q^{9s-2}(q^2+q+2)E_s + q^{18s-12}E_{s-1}.

Now normalize by

Gs(q)=q9(s2)Es(q).G_s(q)=q^{-9\binom{s}{2}}E_s(q).

The recurrence becomes

Gs+1=(1+q1+2q2)Gs+q3Gs1.G_{s+1} = (1+q^{-1}+2q^{-2})G_s + q^{-3}G_{s-1}.

This is precisely the defining recurrence for Fs+1(q1)F_{s+1}(q^{-1}). The initial values agree:

G1=1=F1(q1),G_1=1=F_1(q^{-1}),

and

E2=q6+2q7+q8+q9=q9F2(q1).E_2=q^6+2q^7+q^8+q^9 =q^9F_2(q^{-1}).

Therefore induction gives

Gs(q)=Fs(q1),G_s(q)=F_s(q^{-1}),

or equivalently

ENs,3(2143)(q)=q9(s2)Fs(q1)EN_{s,3}(2143)(q) = q^{9\binom{s}{2}}F_s(q^{-1})

for every s1s\geq1.

This is exactly Conjecture 7.3 of the original source and exactly the statement of MathDB #333178, so the entry should be marked solved.

0 endorsements
Samuel Schlesinger ·