Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace

Research Article

Fast Topology Inference of Wireless Networks Based on Hawkes Process

Download705 downloads
  • @INPROCEEDINGS{10.4108/eai.27-8-2020.2294338,
        author={Yehui  Song and Jiachen  Sun and Guoru  Ding},
        title={Fast Topology Inference of Wireless Networks Based on  Hawkes Process },
        proceedings={Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace},
        publisher={EAI},
        proceedings_a={MOBIMEDIA},
        year={2020},
        month={11},
        keywords={fast topology inference wireless networks hawkes process effectiveness},
        doi={10.4108/eai.27-8-2020.2294338}
    }
    
  • Yehui Song
    Jiachen Sun
    Guoru Ding
    Year: 2020
    Fast Topology Inference of Wireless Networks Based on Hawkes Process
    MOBIMEDIA
    EAI
    DOI: 10.4108/eai.27-8-2020.2294338
Yehui Song1, Jiachen Sun1, Guoru Ding1,*
  • 1: College of Communications Engineering, Army Engineering University, Nanjing 210007, China
*Contact email: dr.guoru.ding@ieee.org

Abstract

Based on the reasonable equivalent assumption of communication events, using the Hawkes process to model the information interaction in wireless communication networks is an emerging direction in the field of non-cooperative topology inference. At present, topology inference algorithms based on the Hawkes process mainly use a fixed sample size for inference, considering only its reliability, but not regarding its effectiveness. In this paper, we consider introducing a sample size as a new performance indicator. For small sample size scenarios in wireless networks, a kind of fast topology inference algorithm is proposed, which uniformly represents parameters belonging to different dimensions, and thoroughly mines topological information from different batches to increase the speed and effectiveness of inference. Experimental simulations show that compared with the existing algorithm, our algorithm has better performance in small sample size scenarios.