Capacity Maximization for Variable-Width WLANs: A Game-Theoretic Approach

This paper investigates non-overlapping variable-width channel allocation for cooperative access points (APs) in multiple collision domains with the goal of maximizing the total capacity of wireless local network (WLAN). Due to the complexity of finding an optimal allocation, this paper considers it from a game-theoretic perspective. First, the problem of variable-width channel allocation is formulated as an identical interest game and the existence of pure Nash Equilibrium (NE) is investigated. Then a decentralized learning-based total capacity maximization algorithm (LTCMA) is designed for APs to achieve an optimal allocation. To analyze the fairness property of the optimal allocation, a game-theoretic fairness analysis model is developed. With this model, this paper shows that the fairness is usually acceptable for a WLAN in which every client is rational and free to associate itself with any APs. Finally, the numerical results verify the effectiveness of LTCMA and the fairness of the optimal allocation.