Two pairs of families of polyhedral norms versus \(\ell _p\)-norms: proximity and applications in optimization (Q263209)

This article considers four novel families of polyhedral norms, each with a single parameter, which are based on the CVaR, the \(\ell_1\) and the \(\ell_\infty\) norms. As these norms can be represented using linear programming, the aim is to compare them with the \(\ell_p\) norm and study their potential use in optimization. The paper begins with an overview of the literature and the necessary preliminary definitions. This is followed by the detailed presentation of the four norms, their properties and the relation with each other. In the fourth section, the authors study the use of these norms in several optimization problems and their performance is evaluated by a series of randomly generated numerical experiments.