提出求解非线性规划问题的一类修正滤子算法.对目标函数和约束违反度函数均设置斜边界作为充分下降条件,相比初始的滤子条件能够改善最优性和可行性,提高了收敛速度.在一般条件下,该算法具有全局收敛性,数值试验也显示出较好的结果.
A modified filter algorithm is proposed for nonlinear programming problem.Slanting envelops have been set for both objective function and constraint violation function.This can improve optimality and feasibility,and provide more sufficient reduction conditions.Under some mild conditions,the algorithm has the property of global convergence.Some numerical tests show the efficiency of our method.