Random permanents and symmetric statistics (Q1338485): Difference between revisions

The theory of permanents of matrices is an interesting and promising approach to matrix algebra. Remember that the determinant of a quadratic matrix is the well-known sum of products of matrix-elements taken one by one from each row and each column with a sign-factor. Now so-called random permanents of rectangular matrices \(A^ m_ n = (a_{ij})_{i=1(1) m,j = 1(1)n}\), \(m\leq n\), with random elements \(a_{ij}\) are studied. Here a permanent \(\text{Per} A_ n^ m\) of a matrix \(A^ m_ n\) is defined as a sum of products, \[ \text{Per} A_ n^ m = \sum_{1\leq j_ 1 \neq \cdots \neq j_ m \leq n} \prod^ m_{s=1} a_{ij_ s}. \] The paper shows that random permanents are a useful and powerful tool in stochastics in the theory of limit theorems. For random permanents of finite-dimensional projection matrices new limiting results are obtained in different asymptotic schemes, e.g. in the Bernoulli scheme with Poisson approximation or in the scheme with normal approximation. The principal statistical functionals such as \(U\)- statistics, symmetric statistics and others, are represented as functionals of permanent measures. Here a permanent of a functional matrix is called functional permanent. The authors intend to develop this approach in future.