一般约束非线性规划问题的光滑牛顿法

用改进的光滑NCP函数替代了文1,2]中的弱互补函数,提出了一种新的光滑牛顿法,从而实现了一般约束优化问题的KKT条件到非线性方程组之间的完全等价转化,且将文3]中提出的求解无约束最优化问题的修正BFGS方法加以改进,应用于求解一般的约束最优化问题,避免了计算Hesse矩阵工作量较大的问题,并在一定的条件下证明了该算法的全局收敛性.

A Smooth Newton Method for General Constrained Nonlinear Optimization Problem

In this paper,an improved smooth NCP function replaces the weak complementary functions in the literature 1,2] and a new smooth Newton method is proposed,therefore the KKT conditions of the constraint optimization is transformed into an equivalent nonlinear equations.the Modified BFGS Algorithm proposed in literature 3] is applied to solve the general constrained optimization problem,the main advantages of this method lies that its iterative matrix is positive definite,while avoiding the calculation of th...