The optimal coefﬁcient in the augmented Lagrangian method for the Saddle Point Problem is found. As the criterion the minimum of the condition number of the diagonal block is taken. The application of the commonly used preconditioners requires the proper approximations of the inverse of this block and of the Schur’s complement. The condition number plays the important role in the calculation of such approximations. The result conﬁrms the experimental value of the coefﬁcient commonly used in the augmented Lagrangian technique.
Digital Object Identifier (DOI)
"On Optimal Coefﬁcient in Augmented Lagrangian Method for Saddle Point Problem,"
Applied Mathematics & Information Sciences: Vol. 11
, Article 2.
Available at: https://dc.naturalspublishing.com/amis/vol11/iss3/2