4th International ICST Conference on Wireless Internet

Research Article

Channel assignment with partially overlapping channels in wireless mesh networks

Download593 downloads
  • @INPROCEEDINGS{10.4108/ICST.WICON2008.4907,
        author={Yong Ding and Yi Huang and Guokai Zeng and Li Xiao},
        title={Channel assignment with partially overlapping channels in wireless mesh networks},
        proceedings={4th International ICST Conference on Wireless Internet},
        publisher={ICST},
        proceedings_a={WICON},
        year={2010},
        month={5},
        keywords={Wireless Mesh Networks Partially Overlapping Channel Channel Assignment Genetic Algorithm.},
        doi={10.4108/ICST.WICON2008.4907}
    }
    
  • Yong Ding
    Yi Huang
    Guokai Zeng
    Li Xiao
    Year: 2010
    Channel assignment with partially overlapping channels in wireless mesh networks
    WICON
    ICST
    DOI: 10.4108/ICST.WICON2008.4907
Yong Ding1,*, Yi Huang1,*, Guokai Zeng1,*, Li Xiao1,*
  • 1: Department of CSE, Michigan State University
*Contact email: dingyong@cse.msu.edu, huangyi7@cse.msu.edu, zengguok@cse.msu.edu, lxiao@cse.msu.edu

Abstract

Many efforts have been devoted to maximizing the network throughput with limited channel resources in multi-radio multi-channel wireless mesh networks. It has been believed that the limited spectrum resource can be fully exploited by utilizing partially overlapping channels in addition to non-overlapping channels in 802.11b/g networks. However, there are only few studies of channel assignment algorithms for partially overlapping channels. In this paper, an extension to the traditional conflict graph model, weighted conflict graph, is proposed to model the interference between wireless links more accurately. Based on this model, we first present a greedy algorithm for partially overlapping channel assignment, and then propose a novel genetic algorithm, which has the potential to obtain better solutions. Through evaluation, we demonstrate that the network performance can be dramatically improved by properly utilizing the partially overlapping channels. In addition, the genetic algorithm outperforms the greedy algorithm in mitigating the interference within the network and therefore leads to higher network throughput.