Fenchel-Lagrange duality versus geometric duality in convex optimization (Q2370031)

In this paper a new duality theory is presented for treating convex optimization problems. It is constructed by means of a perturbation theory and is a combination between the well-known dual problems due to Lagrange and Fenchel. The corresponding strong duality is achieved under weak sufficient conditions and optimality conditions are derived. This duality theory contains as a particular case the geometric duality used by Scott and Jefferson. The authors show that if the new approach is applied to these particular problems, then the strong duality property and the corresponding optimality conditions can be derived under weaker sufficient conditions than the ones obtained by Scott and Jefferson. Finally it is proven that posynomial geometric programming is a special case of the Fenchel-Lagrange duality.