Scarf, Herbert E.; Kannan, R.; Lovasz, Laszlo - Cowles Foundation for Research in Economics, Yale University - 1988
Let A be a real matrix of size (n+d+1)xn. We assume that all n x n submatrices of A are non-singular and define the condition number C = C(A) to be the ratio of the largest n x n subdeterminant of A to the smallest in absolute value. In addition we assume that there is a positive vector pi such...
