Periodicity of Y-systems and flat connections (Q1034789): Difference between revisions

In this interesting paper the author considers an \(Y\)-system \[ Y^WY^E={(1+Y^N)(1+Y^S)\over (1+1/Y_{+})(1+1/Y_{-})}, \] where \(Y^A(\mathbf{m})\) (\(A=W,E,N,S\)), \(Y_{+}(\mathbf{m})\), \(Y_{+}(\mathbf{m})\) are some shift complex valued operators for \(\mathbf{m}\)\( =(n,i,j)\in\Lambda \), \(\Lambda =\mathbb Z\times \mathbb Z\times\mathbb Z\) is a \(3\)-dimensional lattice. Actually, the lattice \(\Lambda \) is a disjoint union of lattices. The author proves existence of periodicity of the system \(A_k\times A_r\) using an interpretation of one as a system of flat connections on a graph. Here the periodicity property appears as an identity among representations of a matrix as products of two-diagonal matrices. The present proof is original, and it differs from one given by \textit{A. Volkov} [On Zamolodchikov's periodicity conjecture for \(Y\)-systems. Preprint, \url{arXiv:hep-th/0606094}] therefore it would be of interest for the reader.