序列二次规划方法为变分不等式问题提供的全局误差界

A global error bound with SQP method for variational inequality problem

针对变分不等式问题,利用序列二次规划方法,定义了一个价值函数.在强单调的条件下,利用价值函数,为变分不等式问题的可行解与最优解之间的距离提供了一个全局误差界.

In this paper, we define a cost function by utilizing sequential quadratic programming method for variational inequality problems. Under the strong monotonicity assumption, the cost function can provide a global error bound for the distance between the feasible point to the optimal solution set.