A modified project variable metric algorithm for equality constrained optimization 等式约束优化一个修正的投影变尺度法
This paper gives a new class of variable metric algorithm for nonlinear programing with equality and inequality constraint conditions, and proves that the algorithm in one-step superlinear convergence when the objective function and constraint functions are uniformly convex. 对带有等式和不等式约束条件的非线性规划给出一类新的变尺度算法,并证明当目标函数和约束函数一致凸时,该算法是一步超线性收敛的。