Computational experience with a primal-dual interior-point method for smooth convex programming

This paper presents some computational experience with a primal-dual interior-point method for smooth convex programming. The algorithm is a direct extension of a primal-dual interior-point method for linear programming. In particular, it does not involve any line search, a definite advantage when function evaluations are costly. The test problems are geometric programming problems and randomly generated convex quadratically constrained problems. We study the influence of the size of the problem (number of inequality constraints and numbers of free variables) on the total number of iterations. Keywords: Interior-point methodPrimal-dual algorithmConvex programmingComputational experience Additional informationNotes on contributors J.P. Vial This research has been supported by the Fonds National de la Recherche Scientifique, grant #122643489