一般约束优化含等式约束问题的鲁棒信赖域算法

摘要

For general constrained optimization, a robust trust region algorithm for inequality constrained optimization problem is extended to general constrained optimization, by using a penalty function method to change general constrained optimization into parametric programming problem. The results show that the perfect characters of the algorithm are retained. Some global convergence results are proved and local superlinear convergence of the algorithm is obtained under standard conditions.%针对一般约束优化问题进行了研究.利用引入罚函数将一般约束问题转化为一个只含不等式约束的的参数规划问题的技巧,将不等式约束优化问题的一个鲁棒信赖域算法扩展到一般约束优化问题中,并保留了算法的良好性质;同时,在一定条件下,得到了算法的全局收敛和超线性收敛.