提出一种新的QP-free方法解变分不等式问题.通过光滑化的Fischer-Burmeister函数,把变分不等式的KKT优化条件转换为一个简单的约束优化问题,并给出了解这个约束优化问题的迭代算法.这个方法的主要优点是:①能够解任意的变分不等式问题;②每步迭代只需解一个线性方程组;③算法是全局收敛的,在一定条件下是超线性收敛的.数值试验结果表明,这个算法是有效的.
A new QP-free method is proposed for the solution of variational inequalities problem. By the smoothing Fischer-Burmeister functions, a simple constrained optimization reformulation of the KKT optimality conditions is derived from variational inequalities. Based on the reformulation, an iterative method is presented. The main properties of this method are: It is well defined for an arbitrary variational inequalities problem; it has to solve just one linear system of equations at each iteration; it is globally convergent; and it is locally superlinerly convergent under certain conditions. Preliminary numerical results indicate that this method is quite promising.