A CLASS OF ALGORITHMS USING SUCCESSIVE QUADRATIC PROGRAMMING WITH PERTURBATIONS AND ITS GLOBAL CONVERGENCE
一种扰动的序列二次规划算法及其全局收敛性

In this paper,a class of successive quadratic programming algorithms for NP is proposed.Instead of exact linear approximations in subproblems like Han's linear approximations withperturbed items are used.To obtain global convergence,a new function for choosing the step-length is developed.