Swarm intelligence based centralized clustering: a novel solution

Abstract

Recent evolutions of MEMS technology, digital electronics, and wireless communication technologies have made smart environments possible; especially through the apparition or incorporation of sensors. Despite being small-sized, the sensors have paved the way for data collection in the environments where they are applied; including luminosity, gas presence, water content, humidity, pressure, and temperature. This research aims to pragmatically optimize Wireless Sensor Network Localization and Network Coverage issues using nature-inspired algorithms. The specific objective is to establish an optimal nature-inspired algorithm and comparing it with other algorithms regarding the capacity to achieve manufacturing optimization in large WSNs, especially in relation to the parameters of high scalability, data delivery rate, and low-energy consumption. Also, the study seeks to determine the extent to which swarm intelligence (SI)-based centralized clustering solutions (optimal nature-inspired algorithms), compared to other approaches, might optimize the WSN features of localization and network coverage. To determine the solutions’ performance, the study involved three scenarios. In scenario 1, the network operational time and the stability of SI-based algorithms were investigated for large WSNs that had the minimum heterogeneity. Imperative to note is that the WSNs on focus had different numbers of nodes, which included 500, 300, and 100. In scenario 2, the motivation was to investigate the SI-based WSN protocols in relation to the parameters of packet delivery, energy conception, and network lifetime for large WSNs. In scenario 3, the performance of SI-based solutions over large WSNs was compared to that which had been reported previously for other algorithms; with the target parameters of comparison being attributes such as packet delivery, energy conception, and network lifetime. From the findings, this study established that SI-based centralized clustering solutions are not only more recent but also exhibit superior performance compared to other algorithms; with the parameters of the amount of data delivered to the BS, energy consumption and scalability on the focus.