Avances en la investigación en ciencias aplicadas Acceso abierto

Abstracto

On Optimal Termination Rule for Primal-Dual Algorithm for Semi-Definite Programming

B.O. Adejo and E. Ogala

In this article, we compare three previous termination rules for primal-dual short step pathfollowing algorithm for semi-definite programming proposed earlier by Monteiro, Adejo and Adejo and Singh independently, which were based on analyses carried out independently by Franklin and Singh et al for Karmarkar’s algorithm for linear programming, Here, we develop a more efficient termination rule which on implementation saves at least 95% iterations over that of Monteiro and at least 8% iterations over that of Adejo and Singh.