Resolving Braess-Like Paradoxes in General Wireless Ad Hoc Networks

Abstract

Coordinating medium access for a massive number of devices is challenging in general wireless ad hoc networks, such as cognitive radio networks and device-to-device (D2D) communication networks. To facilitate medium access, it is often desirable to gather some channel side information (CSI) at the transmitter side such as the channel availability in cognitive radio networks and the channel state information in D2D communication networks. In addition, since it is usually costly to gather the CSI for a large number of devices in a centralized way, distributed medium access control becomes attractive, where each user collects its own CSI to support autonomous and scalable data transmission. However, in the literature, there lacks a systematic way to understand the value of such CSI in the overall network performance. Game theory is often a natural choice for modeling the autonomous and strategic behavior of users. In this paper, we first formulate the multi-band random access game and design a polynomial time algorithm to identify the mixed strategy Nash equilibrium. Then, we propose a systematic framework to incorporate the CSI into the decision making based on the Bayesian game approach. We analyze how the knowledge of the CSI affects the equilibrium strategy and measure the value of this CSI knowledge by the utility difference. We resolve Braess-like paradoxes in general wireless ad hoc networks.