The odd-prime level bound using the last two invariant factors of the walk matrix

Let GG be a controllable graph, let pp be an odd prime factor of detW(G)\det W(G), and let dn1d_{n-1} and dnd_n be the last two invariant factors of the walk matrix W(G)W(G). Write vpv_p for the pp-adic valuation and L(G)L(G) for the level of a rational regular orthogonal matrix associated with GG. Invariant-factor level bound. One should have

vp(L(G))12(vp(dn)+vp(dn1)),v_p(L(G))\leq \frac{1}{2}\bigl(v_p(d_n)+v_p(d_{n-1})\bigr),

and, in particular,

vp(L(G))12vp(detW(G)).v_p(L(G))\leq \frac{1}{2}v_p(\det W(G)).

The proposed bound would remove the technical rank assumption required in the preceding theorem and refine the determinant-based estimate by incorporating the last two invariant factors of the walk matrix.

Sources & referencesView supporting material

Primary source

Wei Wang, Jiaojiao Luo and Li Wang, “On the levels of rational regular orthogonal matrices for generalized cospectral graphs”, arXiv:2602.14213 (2026).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.