The note determines the spectral gap in matroids with parallel elements, indicating specific conditions.
This deposit contains the preprint "Exact spectral gap of the down-up walk on matroids with parallel elements" together with its full replication package. The paper. For a matroid of rank r, the down-up (basis-exchange) walk on its bases has spectral gap at least 1/r, by the log-concavity theorem of Anari, Liu, Oveis Gharan and Vinzant (STOC 2019), and at most 2/r under a mild condition, by a conductance argument of Alev and Lau (STOC 2020). This note determines when the lower end of that window is attained by a degree-1 eigenfunction: a linear observable is an eigenfunction of eigenvalue exactly 1 − 1/r if and only if an explicit functional is constant on bases, and any pair of parallel elements makes it so. Consequently, every matroid with two parallel elements has down-up gap exactly 1/r; for the cographic matroid of a graph, this covers every graph with two edges in series — hence every graph with a degree-two vertex, which includes essentially every sparse graph arising in applications. At the other end of the window, the uniform matroids U(r, 2r−2) are shown to attain the 2/r ceiling exactly, so both ends of the interval [1/r, 2/r] are realized by explicit matroid families. The chain-dependence of these statements is treated with care: four distinct Markov chains from the literature are tabulated on a single running example (the doubled path of Montenegro and Son, STOC 2001, whose mechanism the main theorem explains and generalizes), where their spectral gaps differ by a factor Θ(m). The replication package (ZIP archive) contains the code, seeds, and recorded outputs behind every numerical claim in the paper: exact rational arithmetic verification of the four-chain table, exact enumeration and spectral computation for the control table (including the K₄ edge-subdivision experiment isolating the mechanism), and the kernel computation of the eigenfunction spaces, with an automated test suite and a README mapping each claim of the paper's verification section to the script that produces it.
No takes yet. Share an insight, caveat, or question.
Lucas Damián Alaniz Pintos (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: