Adaptive resource allocation algorithms with QoS support in OFDMA-based WiMAX networks

In Worldwide Interoperability for Microwave Access (WiMAX) the primary concern is Quality of Service (QoS) support which aims to satisfy the diverse service requirements and to guarantee higher data rates allocation for different service classes. However, IEEE 802.16 standard does not specify a bandwidth allocation algorithm to guarantee QoS, this is purposely done in order to allow service providers and vendors to innovate in this area and distinguish their products. The performance benefits of existing solutions in PHY and MAC layers often fall short of providing the QoS support, particularly, it is still experiencing additional access latency and bandwidth allocation disorder where errors occur, that leads flows backlogged. At the same time, mapping the prioritized resources in PHY is become vital to design adaptive resource allocation algorithms that support QoS by way of maximize spectral efficiency, reduce outage probability and efficiently utilize the system resources. The aim of this thesis is to develop a fair and effi- cient packet scheduling and adaptive multiuser frequency-time domain resources allocation algorithms to support QoS for a diverse service class in OFDMA based IEEE 802.16 network. This thesis presents four main contributions for QoS provisioning which are robust, scalable, and can be successfully executed in WiMAX system. The first and second contributions are two slot resource allocation algorithms for OFDMA downlink (DL) scheduling, namely Weighted-rate Adaptive Slot Allocation (WASA) and Feedback Delay-based Slot Allocation (FDSA). The aim is to satisfy QoS requirements for diverse traffic type demands by exploiting available resources in time and frequency domain, and maximize spectral efficiency. These algorithms have been devised with two different approaches. WASA classifies the users based on their weighted-rate factor, which is greater than the minimum requirements, to determine the achievable data rate for each connection in each time-frequency slot. This weighted-rate factor takes into account the achievable data rate along with the QoS requirement to ensure improvements to the system capacity and to guarantee the service type priority for real time over non real-time connections. On the other hand, FDSA allocates the resources based on feedback information delay, which adjusts its traffic in agreement with the feedback to explore the number of slots that must be allocated to the corresponding service type. The purpose of the delay feedback information in real-time service is to utilize available bandwidth efficiently and assign it evenly among the active connections. The Third contribution is an efficient bandwidth allocation algorithm for the uplink transmission called Fuzzy Adaptive Deficit Round Robin (FADRR). FADRR is fully dynamic with fuzzy logic based approach and adaptive deadline-based scheme for various service class traffics in the base station (BS). The algorithm employs fuzzy logic control which is embedded in the scheduler, whereby the function is to control and dynamically update the bandwidth required by the various service classes according to their respective priorities, maximum latency and throughput. FADRR also presents a new adaptive deadline-based approach in order to guarantee a specific maximum latency for real-time connections. The final contribution is Two-Tier hierarchical scheduling algorithm with Enhanced Deficit Round Robin (Two-Tier EDRR) to update and offer new scheduling information to DL and uplink (UL) sub-frame. The main objective is to dynamically allocate the overall bandwidth to DL and UL service classes in such a way that the overall system throughput is optimized without sacrificing their QoS requirements. This is done by assigning the bandwidth fairly among different service classes in a hierarchical structure. The key feature of Two-Tier EDRR is its low-latency queue, in which it allows strict priority queue with delay-sensitive data such as voice to be dequeued and start allocation first before packets in other queues are dequeued. Simulation results indicate that the proposed WASA and FDSA achieve significant performance improvements in terms of spectral efficiency, outage probability, and fairness against the conventional OFDMA-TDMA and MAX-SNR algorithms. WASA obtains higher spectral efficiency in comparison with the OFDMA-TDMA by about 46% and WASA experiences lowest outage probability by about 21%. FDSA attains lower outage probability than OFDMA-TDMA reached by about 51%. In terms of fairness, FDSA outperform MAX-SNR by about 51% and OFDM-TDMA about 71%, respectively. Simulations results show the performance of the proposed FADRR outperforms the conventional MDRR and CDRR schemes. Specifically, it reduces the packet queue delay by 72% and 58% on average in comparison to MDRR and CDRR, respectively. Further, FADRR exhibits better fairness by up to 86% and 61% as compared to MDRR and CDRR, respectively. The throughput for Best Effort (BE) service is maintained at a certain minimum reserved rate, which is still higher by about 27% and 13% in comparison to CDRR and MDRR, respectively. FADRR achieves superior throughput performance for real time Polling Service (rtPS) flows compared to the MDRR and CDRR algorithms by about 18% and 11%, respectively. Two-Tier EDRR provides better performance than the conventional algorithms in terms of end-to-end delay, throughput, and delay jitter for Unsolicited Grant Service (UGS), enhanced real time Polling Service (ertPS), rtPS, non-real time Polling Service (nrtPS) and BE services. It is observed that decreasing delay of real-time packets lead to increase packet delivery ratio, thus enabling the system to show higher throughput.

Preview Text FK 2014 155IR.pdf Download (1MB) | Preview

Actions (login required)

View Item