【摘要】 本文研究了约束优化信赖域法中的线性化约束条件在信赖域内无解的问题.利用一种基于增广Lagrange函数的方法.获得了一个改进的约束优化的信赖域法.该法的线性化约束条件在信赖内有解,并且具有全局收敛性和超线性收敛性.更多还原

【Abstract】 In this paper we study the problem that linearized constraint is infeasible in trust region of trust region algorithm for equality constrained optimization. Using a method based on the augmented Lagrangian function,we get an improved trust region algorithm for equality constrained optimization. Linearized constraint of the improve trust region algorithm is feaible in trust region. Moreover,its global and superlinear convergence ara ensured.更多还原