| Article ID: | iaor19921427 |
| Country: | United Kingdom |
| Volume: | 29 |
| Start Page Number: | 27 |
| End Page Number: | 34 |
| Publication Date: | Oct 1989 |
| Journal: | USSR Computational Mathematics and Mathematical Physics |
| Authors: | Novikova N.M. |
| Keywords: | optimization |
A numerical method is proposed for finding a stochastic minimax saddle point when there is an infinite number of constraints. The method uses a combination of the penalty method and a Uzawa-type stochastic quasigradient method. The parameters of the iterative regularization of the method are coordinated for sufficiently weak regularity conditions on the constraints.