Dynamic Topology Identification of Wireless Communication Networks Based on Hawkes Process

Abstract

ABSTRACT: In the non-cooperative communication network environment, dynamic topology identification of the wireless communication network is a crucial task, because the non-cooperative and dynamic nature of the network poses challenges to topology identification. Information interaction modeling based on the Hawkes process is an emerging direction for non-cooperative network topology identification. The majority of prior topology identification methods based on the Hawkes process only considered static time-invariant topology identification problems, without considering the dynamic nature of the network topology. To tackle this issue, we propose a dynamic topology identification algorithm based on the dynamic window mechanism that is effective in inferencing dynamic network topology. The main contribution of this paper is to model communication events in dynamic networks as multidimensional Hawkes process models. On this basis, combined with the dynamic window mechanism, the Expectation-Maximum method iteratively optimizes the proxy function to identify the topology of the dynamic network. Finally, the identification accuracy performance of link (98.75% accuracy) and non-link (80.11% accuracy) is verified by experiments.