Answers are grounded in this paper and cite their sources.
Let P∈ Z[x]. We study when nonzero integers x₁,x₂,…, uniformly bounded in absolute value, can satisfy \[ ∏ₖ₌₁^n(k+P(k)/x_k)=n! \] for infinitely many n. A degree obstruction forces P≤1, and the two homogeneous affine cases are solved explicitly. For a genuinely mixed affine polynomial $P(x)=ax+b$, we develop a periodic block-rescaling construction governed by a permutation σ∈ SL, leading to the resonance polynomial \[ R_σ(t)=∏ᵣ₌₁^L(1-td_r)-1, d_r=σ(r)-r. \] We isolate the additional integrality and nonvanishing requirements by introducing the notion of an admissible resonance. We show that $0$ is always a double root, establish a reflection symmetry, and reformulate R_σ in terms of elementary symmetric functions of the displacement vector. An exhaustive exact search through block length $9$ identifies the first rational resonances and shows that $L=2,4,7$ are barren, while a partial search at length $10$ finds further examples. We prove an exact direct-sum law for resonance polynomials and use coprime witness lengths to obtain cofinite realization theorems: a slope realized at two coprime admissible block lengths is realized at every sufficiently large length, with one fixed admissible pair $(a,b)$. We give a closed-form proof that block length $4$ is barren, and show that the master construction cannot produce a mixed affine example when b=±1. The paper concludes with conjectures and open questions concerning the full bounded sequence problem beyond the periodic block-rescaling framework.
No takes yet. Share an insight, caveat, or question.
Mohammed Aassila (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: