Expansion in terms of power of a small parameter of the maximum rank distribution of a random Boolean matrix (Q1407105)

The author considers an \(N\times n\) matrix with \(N\geq n\) over a two element field \(GF(2)\). The entries of the matrix are considered as random variables with their distribution depending on a parameter \(\varepsilon\): \[ P\{a_{ij}=0\}=1-P\{a_{ij}=1\}= \tfrac 12(1+\varepsilon x_{ij}), \] for some numbers \(x_{ij}\). The author studies the function \(\chi(A)\), which is 1 whenever the \(n\) columns of \(A\) are linearly independent over \(GF(2)\), and \(0\) otherwise, and finds a power expansion of \(P\{\chi(A)=1\}\) in terms of \(\varepsilon\) and certain coefficients \(f^{(s)}\). Using results from her previous paper [ Theory Stoch. Process. 6(22), 122-126 (2000; Zbl 0973.60016)] she determines exactly the first three coefficients and finds estimates for the rest.