A matrix partition problem (Q796609): Difference between revisions

Let \({\mathcal P}_ n\) be the set of positive definite semi-integral symmetric \(n\times n\) matrices. Any expression of \(T\in {\mathcal P}_ n\) as a sum \(T=T_ 1+T_ 2+...+T_ k\) where \(T_ i\in {\mathcal P}_ n\) (1\(\leq i\leq k)\) is a partition of T. Two partitions of T are the same if the second is obtained by a rearrangement of the constituents in the first. p(T) denotes the number of all different partitions of T. In the paper under review the author introduces the notion of an additively prime element \(T\in {\mathcal P}_ n\) and obtains the generating function \[ F(Z)=\prod_{T\in {\mathcal P}_ n}(1-e^{2\pi i\sigma(TZ)})^{- 1}=1+\sum_{T\in {\mathcal P}_ n}p(T)e^{2\pi i\sigma(TZ)}, \] where \(Z\in {\mathcal H}_ n\) (\({\mathcal H}_ n\) is the Siegel upper-half space of degree n) and \(\sigma\) is the trace of the matrix. He also proves the analyticity of F(Z) in the domain \({\mathcal K}=\{Z=X+iY\in {\mathcal H}_ n| Y\in {\mathcal R}\},\) where \({\mathcal R}\) is the Minkowski reduced domain in the space of positive definite symmetric real matrices of degree n, and considers some elementary cases and numerical examples.