A new self-adaptive alternating direction method for variational inequality problems with linear equality and inequality constraints

In this paper, we present a new self-adaptive alternating direction method for solving a class of variational inequality problems with both linear equality and inequality constraints without the need to add any extra slack variables. The method is simple because it needs only to perform some projections and function evaluations. In addition, to further enhance its efficiency, we adopt a self-adaptive strategy to adjust parameter μ at each iteration. Convergence of the proposed method is proved under certain conditions. Numerical experience illustrates the efficiency of the new method. © 2010 Korean Society for Computational and Applied Mathematics.