Skip to main content
QUICK REVIEW

[Paper Review] Ohno-type relation for finite multiple zeta values

Kojiro Oyama|arXiv (Cornell University)|Jun 2, 2015
Advanced Mathematical Identities5 references4 citations
TL;DR

This paper proves an Ohno-type relation for finite multiple zeta values (FMZVs), confirming a conjecture by Kaneko. Using algebraic structures in the non-commutative polynomial algebra ℋ¹ and properties of Hoffman's dual, the authors establish a duality between sums of FMZVs under index shifts, leading to an alternative proof of the sum formula for FMZVs first proved by Saito and Wakabayashi.

ABSTRACT

Ohno's relation is a well-known relation among multiple zeta values.In this paper, we prove Ohno-type relation for finite multiple zeta values, which is conjectured by Kaneko.As a corollary, we give an alternative proof of the sum formula for finite multiple zeta values, which was first proved by Saito and Wakabayashi.

Motivation & Objective

  • To prove an Ohno-type relation for finite multiple zeta values (FMZVs), confirming a conjecture by Kaneko.
  • To establish a duality between sums of FMZVs under index shifts, linking them to Hoffman's dual indices.
  • To provide an alternative algebraic proof of the sum formula for FMZVs, previously established by Saito and Wakabayashi.
  • To extend the framework of Ohno's relation to the finite zeta setting using the algebraic structure of ℋ¹ and the map T.
  • To demonstrate that the sum formula for FMZVs follows as a corollary from the main duality relation.

Proposed method

  • Define the finite multiple zeta value ζₐ(k) as an element of the algebra 𝒜 = (∏ₚ ℤ/ₚℤ)/(⊕ₚ ℤ/ₚℤ), using p-adic reductions of harmonic sums.
  • Use the non-commutative polynomial algebra ℋ¹ = ℚ⟨x,y⟩y, with generators zₖ = xᵏ⁻¹y, to represent FMZVs via the map Zₐ: ℋ¹ → 𝒜.
  • Employ the harmonic product * and shuffle product X on ℋ¹, with Zₐ preserving the harmonic product and transforming the shuffle product via sign and reversal.
  • Introduce the map T: ℋ¹ → ℋ¹ that implements Hoffman's dual on indices, satisfying Zₐ(T(w)) = ζₐ(k∨) when Zₐ(w) = ζₐ(k).
  • Apply the identity involving the generating function 1/(1−yu)*w = 1/(1−yu) X Δᵤ(w) to derive the main duality relation.
  • Use inclusion-exclusion on the number of non-zero components in exponent vectors to equate sums over shifted indices, proving the main theorem via combinatorial counting.

Experimental results

Research questions

  • RQ1Does an Ohno-type relation hold for finite multiple zeta values, as conjectured by Kaneko?
  • RQ2Can the sum formula for finite multiple zeta values be derived algebraically from a more general duality relation?
  • RQ3How does Hoffman's duality on indices manifest in the finite zeta setting through the algebraic structure of ℋ¹?
  • RQ4What is the precise relationship between sums of FMZVs under index shifts and their duals under the harmonic product?
  • RQ5Can the generating function approach using Δᵤ and the shuffle product be used to prove identities in the finite zeta setting?

Key findings

  • The main theorem establishes a duality: ∑_{|e|=n} ζₐ(k₁+e₁,…,kᵣ+eᵣ) = ∑_{|e′|=n} ζₐ((k′₁+e′₁,…,k′ₛ+e′ₛ)∨), where (k′₁,…,k′ₛ) = (k₁,…,kᵣ)∨.
  • The proof confirms Kaneko’s conjecture on the Ohno-type relation for finite multiple zeta values.
  • The sum formula for FMZVs is recovered as a corollary, with the right-hand side expressed as (−1)ⁱ⁻¹(ℬₖ₋₁ⁱ⁻¹ + (−1)ʳℬₖ₋₁ʳ⁻ⁱ)(Bₚ₋ₖ/k mod p)ₚ.
  • The identity holds for all k,r,i ≥ 1 with 1 ≤ i ≤ r ≤ k−1, and the result is consistent for both even and odd k due to properties of Bernoulli numbers.
  • The coefficient of the sum formula is shown to be (−1)ⁱ⁻¹(−(−1)ᵏ⁺¹ℬₖ₋₁ⁱ⁻¹ + (−1)ʳℬₖ₋₁ʳ⁻ⁱ), which matches the known sum formula under p ≥ k+3.
  • The proof relies on inclusion-exclusion over non-zero components in exponent vectors, showing each index appears exactly once in the alternating sum.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.