Impact of Fairness Models on Resource Allocation in a QoS Constrained Competitive Cognitive Radio Network

Abstract

We consider a multi-channel cognitive radio network (CRN) where multiple secondary users share a single channel and multiple channels are simultaneously used by a single secondary user (SU) to satisfy their rate requirements. In this competitive CRN, our interest is in determining optimal power and rate distribution choices for each SU while maintaining fairness in of across all SUs. Unlike prior approaches that focus on resource allocation based on instantaneous quality of service (QoS), our approach to fairness encompasses both current and prior history of user experience with respect to QoS (collectively referred to as of experience). Specifically, we quantify user experience over time by introducing dynamic fairness weights for each SU in the resource allocation framework. The dynamics of the weights are governed by social behavioral models. In this paper, we study the effect of Homo Egualis (HE), Homo Parochius (HP) and Homo Reciprocan (HR) models. HE or HR based fairness weights in the resource allocation scheme result in a proportionate resource allocation with respect to users demand. Hence, with respect to unweighted allocation scheme (allocation scheme based on instantaneous QoS), the fairness weights in the resource allocation scheme cause some users to sacrifice resources and some users to gain resources to maintain a balance in allocated resources. HP based fairness weights show that collaboration among users through forming group result to achieve a proportionate resource allocation with respect to demand at least across the users of the collaborating group. Simulation results show that considering dynamic fairness weights in the resource allocation scheme provides a better system level fairness index (based on Jain fairness index (1)) relative to the unweighted allocation scheme.. Keywords-Competitive CRN; Quality of Experience; Fairness; Jain fairness index; Home Egualis; Homo Parochius; Homo Reciprocan.