A lightweight heterogeneous network clustering algorithm based on edge computing for 5G

• Published in: Wireless networks Vol. 26; no. 3; pp. 1631 - 1641
• Main Authors: Du, Ruizhong, Liu, Yan, Liu, Liqun, Du, Wenpeng
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.04.2020; Springer Nature B.V
• Subjects: 5G mobile communication; Algorithms; Batteries; Clustering; Communications Engineering; Communications traffic; Computer Communication Networks; Computer simulation; Edge computing; Elections; Electrical Engineering; Electronic devices; Energy consumption; Engineering; IT in Business; Lightweight; Networks; Radio equipment; Residual energy; Wireless networks; 5G; Energy efficiency; Internet of Things; Clustering algorithms; Edge computing
• ISSN: 1022-0038; 1572-8196
• DOI: 10.1007/s11276-019-02144-x

Edge computing is a promising paradigm to provide computing capabilities in close proximity to mobile devices in fifth-generation (5G) networks. However, most wireless sensor devices connected to the 5G network have limited battery life, and how to effectively reduce energy consumption and extend the network life cycle has become one of the hot problems in current research. Due to this motivation, an improved Stable Election Protocol (SEP), named Lightweight in Edge Computing-SEP (LEC-SEP) is proposed. LEC-SEP algorithm considers the heterogeneity of the initial energy of the nodes and the cluster head election is determined by the probability that the relative level of the initial energy and the residual energy. According to the influence of the number of cluster heads, the optimal clustering number is calculated to balance the network traffic. At the same time, the location of the base station is redefined to facilitate adding the edge server, which can store the data aggregated and fused by base station, providing powerful and real-time storage and computing power to effectively offload the pressure of the central cloud. The simulation results show that the energy consumption is well distributed in the proposed algorithm, and LEC-SEP algorithm achieves a longer stabilization period in the network than other typical clustering algorithms. The network life of LEC-SEP improved by 8.17% and 20.34% in comparison with the P-SEP algorithm and the IDEEC algorithm respectively.