On information theoretic games for interference networks

The Nash equilibrium region for a 2-user game was defined and characterized first for a linear deterministic channel and then for a Gaussian channel. Challenges in extending this understanding to some special K-user cases have also been explored. In this paper, we study two indices which compare the performance (sum-rate) of the `best´ and the `worst´ Nash equilibria to the optimum (sum-capacity) and reflect the `price of stability´ and the `price of anarchy´, respectively. These indices are evaluated for the 2-user and some special K-user linear deterministic interference channels. We further investigate the impact on these indices of changing the payoff functions of each user to include a cost of transmission.