Home

In this website you can find the matrices Q used in the computational experiments described in the manuscript "On the Solution of Nonconvex Cardinality Boolean Quadratic Programming problems. A computational study” by Ricardo M. Lima and Ignacio E. Grossmann".

The matrices Q were generated using a random number generation function with a uniform distribution for a given lower bound, LB, upper bound, UB, and sparsity density, SD. 16 families of matrices Q are built for Q(LB,UB,SD) = {(−100, 100, SD), (−1, 1, SD), (0, 1, SD), (0, 100, SD), SD ∈ {0.1, 0.5, 0.75, 1}}, and for each family, five matrices are generated, resulting in 80 instances of the matrix Q spanning the interval [-100, 100].

A compressed file with the matrices can be found here.