A dynamical neural network for solving stochastic two-player zero-sum games
Résumé
This paper presents a dynamical neural network approach to solve stochastic two-players zero-sum game problems. The original problem is first transformed into an equivalent convex second-order cone programming problem. We develop a dynamical neural network model to solve the problem, where the model's equilibrium point corresponds to the optimal solution of the game problem. Further, we use a Lyapunov function to show that the equilibrium point of the neural network is globally asymptotically stable. Numerical results are given to show the performance of our approach.
Origine : Fichiers produits par l'(les) auteur(s)