最优化广义次可行方向法的统一模型及收敛性

An unified model of generalized subfeasible direction methods for optimization and its convergence

讨论一般约束最优化问题 ,首先将原问题转化为只带不等式与线性等式约束的扩充优化问题 ,然后利用次可行方向法思想及梯度投影法 ,通过扩充问题建立原问题的一类十分一般的求解模型 .分析了模型在原问题的

In this paper, the general constrained optimization problems are discuussed. At first, they are changed into the expansive optimization problems holding with only linear equality and inequality constraints, then a more general algorithm model for solving the primial problems is presented by using the expansive problems as well as the ideas of subfeasible direction methods and gradient projection. The global convergence of this model is discussed.