This thesis addresses the problem of finding the extreme volumes of quasi-copulas, which are generalized forms of copulas used to model dependencies between multidimensional random variables. The extreme volumes of quasi-copulas provide a quantitative measure of how much quasi-copulas can differ from copulas. We focus on finding extreme volumes for higher dimensions using the approach of linear programming. The derived linear program is solved using the simplex method and the interior-point method in Mathematica and Matlab, and the obtained results are compared. It turns out that the conjecture for calculating the lower bound of the volume in the 2023 article does not hold. The extreme volumes, obtained using the simplex method in Mathematica, are in the form of rational numbers and may contribute to the possibility of determining an exact formula for the extreme volumes of quasi-copulas.
|
