Answers are grounded in this paper and cite their sources.
Jensen [arXiv:2607.01173] computes the tropical Prym–Brill–Noether locus V^r(Γ,π) of the k-uniform double cover of a loop of loops, proves dim V^r(Γ,π) ≤ g−1−⌈k(r+1)/2⌉ for k 2. We answer both questions in the negative. For g−1 ≥ N(k,r), the dimension equals g−1−N(k,r), where N(k,r) is the least number of k-uniform blocks whose Demazure product is the longest element w_0 of S_(r+1). By exhaustive computation N(3,5)=10>9, and at (k,r)=(4,5) the irredundant words have lengths 12 and 13, so V^r(Γ,π) is not pure for g−1 ≥ 13. We compute N(k,r) for all 2 ≤ k ≤ r ≤ 10 and in a larger range for odd k, always finding N(k,r)=(k+1)r/2 for odd k ≤ r, and conjecture this in general. We prove the upper bound N(k,r) ≤ (k+1)r/2 for every odd k ≥ 3 and r ≥ 2: by a 0-1 reduction the schedule c_t ≡ 2t (mod k) becomes a deterministic exclusion process whose hop times satisfy a constant-weight last-passage recursion, solved in closed form. For period-k schedules, which expand to repetitions of Coxeter elements, we prove that for r ≥ k no odd-k schedule sorts in fewer than (k+1)r/2 blocks; in the block-time formulation the obstruction is the parity of a winding number. For schedules of period 2k, and of period nk with n ≤ k+1 in which every class fires equally often, we prove the conjectured growth rate (k+1)/2 via maximum cycle means. We also record a structural conjecture on the graded orbit of the block monoid, with computational evidence.
No takes yet. Share an insight, caveat, or question.
Andre Ruiz Loera (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: