A new penalty function method for constrained minimization
Barry W. Kort, Dimitri P. Bertsekas · 1972
During recent years it has been shown that the performance of penalty function methods for constrained minimization can be improved significantly by introducing gradient type iterations for solving the dual problem. In this paper we present a new penalty function algorithm of this type which offers significant advantages over existing schemes for the case of the convex programming problem. The algorithm treats inequality constraints explicitly and can also be used for the solution of general mathematical programming problems.