Indian Journal of Science and Technology
Year: 2015, Volume: 8, Issue: 23, Pages: 1-10
Balamuralikrishna Potti1*, M. V. Subramanyam2 and K. Satya Prasad3
1 JNTUK, Kakinada - 533 003, Andhra Pradesh, India; [email protected]
2 Department of ECE, Santhiram Engineering College, Nandyal, Kurnool - 518 501, Andhra Pradesh, India; [email protected]
3 Department of ECE, University College of Engineering, JNTUK, Kakinada - 533 003, Andhra Pradesh, India; [email protected]
Background/Objectives: The objective of this work is to mitigate starvation in Wireless Mesh Networks (WMNs) being deployed in today’s LAN, WAN and Internet topologies by employing a novel optimization method. Methods/Analysis: The QoS performance of WMNs is severely affected by a problem called starvation where nodes that are one-hop away from the gateway monopolize the channel so that far away nodes get starved of channel access. We propose a hybrid genetic algorithmic approach to mitigate starvation in WMNs by dynamic adjustment of contention window of mesh nodes optimally. In this approach, Genetic Algorithm incorporated with Gravitational Search Algorithm is used. Findings: Simulations are conducted using the proposed method for multimedia traffic with AODV as the routing protocol. The performance of the proposed method is compared with priority-based method, pure GA optimization, Fair Binary Exponential Back-off algorithm (FBEB) and IEEE 802.11. The local search capability of GSA incorporated in our proposed method improves the throughput by 24.64% than priority-based method and by 3.56% than pure GA optimization. In our approach, we observed a significant decrease in end-to-end delay compared to pure GA optimization. Improvement in fairness is found along one-hop, two-hop and three-hop nodes when compared with FBEB. The FBEB algorithm adjusts the CW size by indirectly estimating the traffic in communication medium leading to lesser throughput whereas our proposed method changes the CW size dynamically based on QoS parameters of network nodes leading to improvement in throughput. The proposed method increases throughput by 5.10% than IEEE 802.11 and by 1.25% than FBEB at one-hop. The proposed method increases throughput by 66.67% than IEEE 802.11 and by 22.61% than FBEB at two-hop. Application/Improvement: Our hybrid Genetic optimization method improves QoS performance of Wireless Mesh networks by avoiding throughput imbalances among users and reducing end-to-end delay effectively.
Keywords: Contention Window, Genetic Algorithm, Gravitational Search Algorithm, Starvation, Wireless Mesh Networks
Subscribe now for latest articles and news.