Cohen–Macaulayness classification for cyclic-graph invariant chains

Determine whether, for every integer n≥6 and r≥0, the ideal I_{n+r}:=\mathrm{Inc}(\mathbb{N})_{n,n+r}(I(C_n)) is Cohen–Macaulay if and only if r≥\lfloor(n-4)/2\rfloor and r\ne n-4.

Background

The paper studies chains of edge ideals generated from a graph on [n] by applying all strictly increasing injections into [n+r]. For the cyclic graph C_n, the authors prove that the resulting ideal becomes the edge ideal of the complete graph K_{n+r} whenever r≥n-3, which establishes Cohen–Macaulayness in that range.

Computational and structural evidence suggests that the complete-graph threshold is not optimal. The proposed classification asserts Cohen–Macaulayness beginning already at r=\lfloor(n-4)/2\rfloor, except at the isolated value r=n-4. At that exceptional value, the independent pairs form a disjoint union of antipodal edges, making the independence complex disconnected and therefore preventing Cohen–Macaulayness.

References

For cyclic graphs we prove $\mathrm{Inc}(\mathbb{N}){n,n+r}(I(C_n))=I(K{n+r})$ once $r\ge n-3$, giving Cohen--Macaulayness in this range, and conjecture -- with supporting computational and structural evidence -- a complete classification: Cohen--Macaulayness holds if and only if $r\ge\lfloor(n-4)/2\rfloor$ and $r\ne n-4$.

Cohen--Macaulayness of $\mathrm{Inc}(\mathbb{N})$-Invariant Chains of Edge Ideals  (2609.03566 - Anwar et al., 3 Sep 2026) in Conjecture 3.14, subsection “Complete graphs and cyclic graphs under \mathrm{Inc}(\mathbb{N})”; also stated in the Abstract and Introduction