提出了—个求解非线性互补约束均衡问题的滤子SQP算法.借助Fischer—Burmeister函数把均衡约束转化为一个非光滑方程组,然后利用逐步逼近和分裂思想,给出一个与原问题近似的一般的约束优化.引入滤子思想,避免了罚函数法在选择罚因子上的困难.在适当的条件下证明了算法的全局收敛巨,部分的数值结果表明算法是有效的.
In this paper, a new method of SQP-filter for mathematical programs with non- linear complementarity constraints is proposed. By means of F-B function, the nonlinear complementarity constraints condition is transformed into a nonsmooth equations, and then the constrained optimization problem similar to the original problem is given by the use of successive approximation and decomposition. The difficulty of choosing the penalty param- eter associated with use of penalty functions can be avoided by introducing a new concept of "filter". Under suitable conditions, the global convergence is proved. The limited numerical test shows its efficiency.