摘要
将信赖域方法用于不等式约束问题，在不要求Ｈｅｓｓｉａｎ阵的近似矩阵正定的情况下，证明了算法的整体收敛性，且在一定条件下，证明了算法是二阶收敛的．最后给出了１个计算实例．
This paper presents a trust region-based method for the general nonlinear programmingproblem. The method works by changing the constrains from inequality to equality，then us-ing the trust region method．It is proved that the method is globally convergent without thecondition that Hessian approximation is positively definite，and that under better conditions，the method is quadratically convergent.
