Abstract (100 words or less): |  | We present a new approach to solving
nonlinear programming (NLP) problems in which the strict complementarity condition, constraint qualifications and second-order sufficient conditions for optimality are not necessarily satisfied at a solution. In a local neighborhood of the solution, we construct a modified Lagrange system that has a locally unique regular solution. By applying Newton's method to the Lagrange system, we get a method with a superlinear rate of convergence to the solution of the NLP problem.
The proposed approach can be applied to solving a variety of problems. |