Article ID: | iaor20134038 |
Volume: | 56 |
Issue: | 2 |
Start Page Number: | 373 |
End Page Number: | 397 |
Publication Date: | Jun 2013 |
Journal: | Journal of Global Optimization |
Authors: | Strodiot Jean, Nguyen Van, Nguyen Thi |
Keywords: | line search, Nash equilibrium, extragradient |
Generalized Nash equilibrium problems are important examples of quasi‐equilibrium problems. The aim of this paper is to study a general class of algorithms for solving such problems. The method is a hybrid extragradient method whose second step consists in finding a descent direction for the distance function to the solution set. This is done thanks to a linesearch. Two descent directions are studied and for each one several steplengths are proposed to obtain the next iterate. A general convergence theorem applicable to each algorithm of the class is presented. It is obtained under weak assumptions: the pseudomonotonicity of the equilibrium function and the continuity of the multivalued mapping defining the constraint set of the quasi‐equilibrium problem. Finally some preliminary numerical results are displayed to show the behavior of each algorithm of the class on generalized Nash equilibrium problems.