摘要

As for variational inequality problems, we define the optimal value function of trust region subproblems, study the properties of it. And then under strongly monotonic conditions, we use the optimal value function to provide a global error estimate for the feasible solution.