A two-phase algorithm for a variational inequality formulation of equilibrium problems (Q2442699)

The authors consider a general equilibrium problem, that is to find a point \(x^{*}\in C\) such that \[ f (x^{*}, y) \geq 0 \quad \forall y \in C, \] where \(C\) is a nonempty convex and closed set in a Euclidean space, \(f (x, x)=0\) for every \(x \in C\), and \(f (x, \cdot )\) is convex for each \(x\). They propose an iterative method applied in fact to the equivalent variational inequality based on the Rosen mapping, where the feasible set \(C\) is replaced by its sequential approximations for making projections. The approximation is suggested to be found by the subgradient method with the known optimal value. The main iteration can also be treated as some subgradient procedure with divergent series stepsize rule. The method is a modification of similar previous ones; e.g. [\textit{J. Y.~Bello Cruz} and \textit{A. N. Iusem}, Math. Comput. Simul. (accepted)] and convergent under paramonotonicity type assumptions. Some results of experiments are also reported. However, the authors do not compare the method with the known ones having better theoretical convergence properties. For instance, the corresponding versions of the averaging method [\textit{I. Konnov}, Combined relaxation methods for variational inequalities. Lecture Notes in Economics and Mathematical Systems. 495. Berlin: Springer (2001; Zbl 0982.49009), Ch. 2] have similar expenses per iteration, but are convergent under the (generalized) monotonicity.