一类关于Uzawa-AOR方法的鞍点问题

A Class of Uzawa-AOR Method for Saddle Point Problems

对于大型稀疏鞍点问题,本文研究一类用于求解鞍点问题的Uzawa-AOR方法,我们得出了保证其收敛的迭代方法。实际上,与Uzawa为外迭代和AOR为内迭代的方法相比,新的方法可以被认为是一个不精确的迭代。最后数值算例结果表明,新的迭代方法可以减少每一步的迭代数并且具有更快的收敛速度。

For large sparse saddle point problems, in this paper,we consider a class of Uzawa-AOR method for solving the saddle point problems. We derive conditions for guaranteeing the convergence for the iterative method. Actually, the new method can be considered as an inexact iteration with the Uzawa as the outer iteration and the AOR as the inner iteration. Finally, numerical example shows that the resulting new method leads to less workload per iteration step and fast convergence.