[Paper Review] A short proof of Kontsevich cluster conjecture

Arkady Berenstein, Vladimir Retakh|arXiv (Cornell University)|Nov 1, 2010

Algebraic structures and combinatorial models1 references3 citations

TL;DR

This paper provides a concise, elementary proof of Kontsevich's cluster conjecture, showing that iterated applications of the Kontsevich map $K_r$ on a noncommutative plane yield noncommutative Laurent polynomials in $x$ and $y$ for any positive integers $r_1, r_2$. The proof relies on constructing a noncommutative cluster algebra ${\mathcal{A}}(r_1,r_2)$ generated by recursively defined elements $y_k$, and establishes that all such $y_k$ lie within this algebra, thereby confirming the conjecture via algebraic invariance and recursive relations in a free group algebra.

ABSTRACT

We give an elementary proof of the Kontsevich conjecture that asserts that the iterations of the noncommutative rational map K_r:(x,y)-->(xyx^{-1},(1+y^r)x^{-1}) are given by noncommutative Laurent polynomials.

Motivation & Objective

• To provide a short, elementary proof of Kontsevich's conjecture on the Laurentness of iterated Kontsevich maps on the noncommutative plane.
• To establish that all iterations of the map $K_{r_1}K_{r_2}K_{r_1}\cdots$ yield noncommutative Laurent polynomials in $x$ and $y$ for any $r_1, r_2 \in \mathbb{Z}_{>0}$.
• To define and analyze a purely noncommutative cluster algebra ${\mathcal{A}}(r_1,r_2)$ as the subalgebra generated by recursively defined elements $y_k$ in the division ring of the free group algebra $\mathbb{Q}\langle y_1^\pm, y_2^\pm \rangle$.
• To prove that the algebra ${\mathcal{A}}(r_1,r_2)$ is invariant under index shift, i.e., ${\mathcal{A}}_k = {\mathcal{A}}$ for all $k \in \mathbb{Z}$, thereby ensuring closure of the iteration process.

Proposed method

• Define the Kontsevich map $K_r: (x,y) \mapsto (xyx^{-1}, (1+y^r)x^{-1})$ and consider its iterated application in alternating fashion for $r_1, r_2 \in \mathbb{Z}_{>0}$.
• Introduce the commutator $z = [x,y] = xyx^{-1}y^{-1}$, which remains invariant under the map, and use it to derive recursive relations for $x_k, y_k$.
• Reformulate the dynamics in terms of $y_k$ elements via the recursion $y_{k+1} z y_{k-1} = 1 + y_k^{r_k}$, with $r_k$ alternating between $r_1$ and $r_2$.
• Work within the group algebra $\mathcal{F}_2 = \mathbb{Q}\langle y_1^\pm, y_2^\pm \rangle$, which embeds into a division ring, to define $y_k$ recursively for all $k \in \mathbb{Z}$.
• Define ${\mathcal{A}} = {\mathcal{A}}(r_1,r_2)$ as the subalgebra generated by $y_0, y_1, y_2, y_3, z, z^{-1}$, and prove that $y_k \in {\mathcal{A}}$ for all $k \in \mathbb{Z}$ using a shift-invariance argument.
• Use an anti-automorphism $\sigma$ swapping $y_1 \leftrightarrow y_2$ and $y_k \leftrightarrow y_{3-k}$ to establish symmetry and prove that $y_k \in {\mathcal{A}}_{k+1}$, completing the invariance proof.

Experimental results

Research questions

• RQ1 Does the iterated application of the Kontsevich maps $K_{r_1}$ and $K_{r_2}$ on the noncommutative plane always yield noncommutative Laurent polynomials in $x$ and $y$ for any positive integers $r_1, r_2$?
• RQ2 Can the recursive structure of the map dynamics be captured algebraically within a noncommutative cluster algebra framework?
• RQ3 Is the subalgebra ${\mathcal{A}}(r_1,r_2)$, generated by $y_0, y_1, y_2, y_3, z, z^{-1}$, invariant under index shift, i.e., does ${\mathcal{A}}_k = {\mathcal{A}}$ hold for all $k \in \mathbb{Z}$?
• RQ4 What are the defining relations of the noncommutative cluster algebra ${\mathcal{A}}(r_1,r_2)$, and do they include the key identities such as $y_{k+1} z y_k = y_k y_{k+1}$ and $y_{k+1} z y_{k-1} = 1 + y_k^{r_k}$?

Key findings

• The main result confirms Kontsevich's conjecture: all iterations $\underbrace{\cdots K_{r_1}K_{r_2}K_{r_1}}_{k}(x,y)$ for $k \geq 1$ yield noncommutative Laurent polynomials in $x$ and $y$ for any $r_1, r_2 \in \mathbb{Z}_{>0}$.
• The algebra ${\mathcal{A}}(r_1,r_2)$ is invariant under index shift: ${\mathcal{A}}_k = {\mathcal{A}}$ for all $k \in \mathbb{Z}$, which implies that all $y_k$ lie in ${\mathcal{A}}$.
• The elements $y_k$ satisfy the key identity $y_{k+1} z y_k = y_k y_{k+1}$ for all $k \in \mathbb{Z}$, which is proven by induction using the recursive relation and the Malcev divisibility of the free group algebra.
• A noncommutative version of the cluster algebra is constructed via recursive relations in the free group algebra, and its structure is shown to be closed under the dynamics of the Kontsevich map.
• The anti-automorphism $\sigma$ satisfying $\sigma(y_k) = y_{3-k}$ and $\sigma({\mathcal{A}}_k(r_1,r_2)) = {\mathcal{A}}_{-k}(r_2,r_1)$ is used to prove the inclusion $y_k \in {\mathcal{A}}_{k+1}$, completing the invariance proof.
• The defining relations of ${\mathcal{A}}(r_1,r_2)$ include $y_i y_{i+1} = y_{i+1} z y_i$, $y_{j+1} z y_{j-1} = y_j^{r_j} + 1$, and a complex relation involving $y_3 z y_0 - z y_0 y_3 z = y_2^{r_2-1} y_1^{r_1-1} - z (y_1 z)^{r_1-1} (y_2 z)^{r_2-1}$, which is verified using group algebra identities.