Global Convergence of an Improved Filter Method of Feasible Direction for Inequality Optimization

In this paper, we proposed a global convergence of an improved filter method of feasible direction for solving inequality constrained optimization. At every trial point, it is only necessary to solve one QP sub problem with equality constraints. On contrary the traditional SQP algorithm, we use filter method with the new SQP algorithm, which avoid Maratos effect. Under some reasonable conditions, the global convergence result of our algorithm is presented.