Adaptive hierarchical polling and Markov decision process based CAC for increasing network reward and reducing average delay in IEEE 802.16 WiMAX networks

The IEEE 802.16 standard, namely WiMAX, adopts a polling MAC to support real-time transmissions in wireless metropolitan area networks. Although WiMAX overcomes access contentions by using unicast polling, it suffers from long polling delay and wastes bandwidth used for polling when the polled SSs have no packets to be sent. Additionally, WiMAX supports various service-classes for guaranteeing QoS transmission, but obtains reward only from high-class services because WiMAX doesn't differentiate the priorities of nodes. However, today's network providers have already supported different customer-priorities and service-classes simultaneously for increasing network reward. This paper thus proposes the adaptive hierarchical polling (i.e., AHP) approach that consists of three schemes. First, AHP supports multiple node-priorities with different classes of service flow to maximize network reward. Second, a two-mode pending mechanism is proposed for the BS and SSs to dynamically change any SSs state between the active and idle states. Third, we propose the MDP-based CAC to increase network reward by accepting high-priority nodes and real-time polling services first. Numerical results indicate that the proposed approach outperforms IEEE 802.16 in FRL and polling delay while yielding competitive utilization. Furthermore, the worst-case running time of AHP is analyzed and deduced to O(N . W . R). (C) 2008 Elsevier B.V. All rights reserved.