基于积极集技术求解无约束极大极小问题的摄动SQP方法

Perturbed SQP method with active set technology for unconstrained minimax problems

讨论无约束极大极小(minimax)问题,基于积极集识别技术,结合摄动的序列二次规划(SQP)方法,建立问题的一个数值方法.在相当弱的条件下,算法具有弱全局收敛性,并对算法进行了初步的数值试验.

In this paper, the unconstrained minimax problems are discussed. By combining active set identification technique with perturbed sequential quadratic programming （SQP） method, a numerical method is proposed for the problems. Under quite weak conditions, the proposed algorithm possesses weak global convergence. Some preliminary numerical experiments are carried out.