Generalized Nash equilibrium problems under partial-decision information with biased agents

We consider generalized Nash equilibrium (GNE) problems under a partial-decision information regime, in which each agent typically reconstructs the opponents' strategies through a linear averaging dynamics. In contrast, we consider a state-dependent, nonlinear susceptibility term within the communication mechanism, thereby modelling possible biases on the part of agents in processing information. By including such a term in a relaxed forward-backward iteration scheme, we design a distributed algorithm possessing convergence guarantees to a GNE. Simulation results illustrate how the susceptibility term affects the GNE computation.