Online Distributed Algorithms for Seeking Generalized Nash Equilibria in Dynamic Environments

In this article, we study the distributed generalized Nash equilibrium (GNE) seeking problem of noncooperative games in dynamic environments. Each player game aims to minimize its own time-varying cost function subject a local action set. The sets all players are coupled through shared convex inequality constraint. can only have access function, set constraint, and block communicate with neighbors via connected graph. Moreover, do not prior knowledge their future functions. To address problem, an online algorithm is proposed based on consensus algorithms primal-dual strategy. performance measured by using regrets. Under mild assumptions graphs functions, prove that if deviation variational GNE sequence increases within certain rate, then regrets, as well violation grow sublinearly. A simulation presented demonstrate effectiveness our theoretical results.