A Hierarchical Learning Solution for Anti-Jamming Stackelberg Game With Discrete Power Strategies

This letter investigates the anti-jamming problem with discrete power strategies, and then a Stackelberg game is formulated to model the competitive interactions between the user and jammer. Specifically, the user acts as the leader, whereas the jammer is the follower. Based on their own utilities, the user and jammer select their power strategies and determine their respective optimal strategies. Also, a hierarchical power control algorithm (HPCA) is proposed to obtain the Stackelberg equilibrium, and the asymptotic convergence is analyzed. In addition, we consider the impact of the imperfect information due to the jammer's bounded rationality and inaccurate observation of the user's action. Finally, simulations are conducted to show the effectiveness of the proposed HPCA algorithm, and simulation results demonstrate that the jammer's bounded rationality and limited observation lead to the increase of the user's utility.