利用了序列二次规划来求解非线性规划问题,并且引进滤子的概念。这样做可避免使用罚函数时选择罚函数参数的困难。在算法中,每次迭代分成可行性阶段和最优化阶段。在可行性阶段中,减小不可行性的某种度量:在最优化阶段中,减小目标函数值。在一些较弱的条件下,证明了算法的全局收敛性。
In this paper the solution of nonlinear programming problems by a Sequential Quadratic Programming (SQP) algorithm is considered and a new concept of ‘filter' is introduced. The difficulty of choosing the penalty parameter associated with the use of penalty functions can be avoided. In this algorithm, each iteration is composed of a feasibility phase, which reduces a measure of in.feasibility, and an optimality phase, which reduces the objective function. Based on some assumptions for the feasibility and optimality phases, the global convergence to stationary points will be proved.