Capacity Maximization Using Nash Bargaining in Indoor Optical Multiuser Interference-Limited Poisson Channel

In multiuser optical wireless communications (OWC), the multiuser interference becomes significant as the number of users increases. This would lead to a poor quality of service with a low signal-to-interference-plus-noise-ratio (SINR) or even a failure of service. In this paper, we consider coordination among multiple users to maximize the capacity in a multiple-input single-output (MISO) interference-limited discrete-time Poisson channel from a game-theoretic perspective. With power constraints imposed, the resource allocation problem can be interpreted as a downlink beamforming game. To improve the capacity in a power- and interference-limited channel, we model the problem as a cooperative game and propose an altruistic algorithm using the Nash bargaining, where each self-interested user attempts to maximize its capacity through cooperation. Moreover, to compare the performance with a non-game-theoretic approach, a convex optimization framework is developed. It is shown that the proposed Nash bargaining yields significant performance improvement when compared with the Nash equilibrium in the non-cooperative game. It is also found that in a power- and interference-limited channel, self-interested but cooperative users can provide a significant performance improvement over the convex optimization based non-game-theoretic approach.