AN SQP ALGORITHM OF LINEAR INEQUALITY CONSTRAINTS

In this paper, a kind of optimization problems with linear inequality constraints are discussed, and a sequential quadratic programming feasible descent algorithm for solving the problems is presented. In order to show the algorithm is well defined, we obtained a feasible descent direction by using generalized projection technique. The high-order revised direction which avoids Maratos effect is computed. Under some suitable conditions, the global and superlinear convergence can be obtained.