QoE-driven multi-service resource scheduling strategy in mobile network

Abstract

As quality of experience (QoE) concerns more about users' end-to-end subjective experience than quality of service (QoS), it becomes an important performance metric when designing a resource scheduling algorithm. In this paper, we propose a QoE-driven multi-service resource scheduling (QMRS) algorithm aiming at maximizing the QoE of the whole system. In QMRS, a specific utility model is adopted as a normalized QoE evaluation metric of end users, which is highly universalizable and extensible and of great importance for the newborn service evaluation. We use a greedy algorithm based on utility models for different services to optimize the wireless resource allocation in multi-users mobile network. Compared with the traditional proportional fair (PF) scheduling method, the end users' utility value increases from 0.82 to 0.92 in less users condition. In condition of 45 users, the utility value can increase to 0.56 with QMRS method from 0.26 with PF method. The results validate that the proposed QMRS can guarantee users' QoE in different services with limited wireless resource.