Joint Rate Adaptation and Medium Access in Wireless LANs: a Non-cooperative Game Theoretic Perspective

TL;DR

This paper models non-cooperative WLANs with selfish users adjusting data rates and contention windows, proposing a game-theoretic approach with an adapted Tit-For-Tat strategy to reach efficient equilibria close to optimal system performance.

Contribution

It introduces a non-cooperative game model for joint rate and medium access control in WLANs, with a novel adapted Tit-For-Tat strategy ensuring efficient equilibrium.

Findings

Multiple equilibria exist in the game, but refined to a unique efficient one.

The distributed algorithm converges to an equilibrium near the system optimum.

The approach effectively aligns individual incentives with overall network efficiency.

Abstract

Wireless local area networks (WLANs) based on IEEE 802.11 standards are becoming ubiquitous today and typically support multiple data rates. In such multi-rate WLANs, distributed medium access and rate adaptation are two key elements to achieve efficient radio resource utilization, especially in non-cooperative environments. In this paper, we present an analytical study on the non-cooperative multi-rate WLANs composed of selfish users jointly adjusting their data rate and contention window size at the medium access level to maximize their own throughput, irrespective of the impact of their selfish behaviors on overall system performance. Specifically, we develop an adapted Tit-For-Tat (TFT) strategy to guide the system to an efficient equilibrium in non-cooperative environments. We model the interactions among selfish users under the adapted TFT framework as a non-cooperative joint medium access and rate adaptation game. A systematic analysis is conducted on the structural properties of the game to provide insights on the interaction between rate adaptation and 802.11 medium access control in a competitive setting. We show that the game has multiple equilibria, which, after the equilibrium refinement process that we develop, reduce to a unique efficient equilibrium. We further develop a distributed algorithm to achieve this equilibrium and demonstrate that the equilibrium achieves the performance very close to the system optimum in a social perspective.