A hybrid gazelle optimization and reptile search algorithm for optimal clustering in wireless sensor networks

Abstract

In our modern societies, the wireless sensor network (WSN) is categorized as a smart motivated technology that can be utilized in many work environments and activities to enhance daily life. However, several challenging concerns have been assigned to WSN. The clustering process is a main complex concern and still an open problem in WSN. To support an efficient clustering process, two crucial requirements must be considered, energy management and network lifetime extension, especially in the development of large-scale WSN. The primary objective of this article is to introduce a new meta-heuristic algorithm, denoted as the hybrid gazelle optimization and reptile search algorithm (HGORSA), which optimizes cluster head selection in WSNs. In the proposed algorithm, the mathematical models for the exploration and exploitation phases of the traditional gazelle optimization algorithm (GOA) are enhanced by integrating the hunting operator, reduction function, and predator cumulative effect operators from the traditional RSA. These modifications improve the balance between diversification and intensification processes, effectively addressing two key clustering requirements mentioned above. At the same time, they also positively impact the overall performance evaluation of the WSN. Various simulation scenarios are designed to evaluate the performance of the proposed HGORSA in different network configurations. First, the main experiment was conducted with 300 sensor nodes (SNs). The experimental results then analyzed to assess the effectiveness of the proposed algorithm under different conditions against six state-of-the-art meta-heuristic algorithms. Based on simulation outputs, HGORSA demonstrated superior performance compared to particle swarm optimization, grey Wolf optimizer, sperm swarm optimization, chernobyl disaster optimizer, gazelle optimization algorithm and reptile search algorithm. Specifically, HGORSA achieved percentage improvements in terms of stability period (37.3%, 49.6%, 46.8%, 55.3%, 19.1%, and 34.4%, respectively), energy consumption (10.8%, 10.5%, 9.6%, 8.6%, 8.3%, and 3.5%, respectively), network lifetime (44.5%, 40.8%, 23.8%, 16.8%, 9.3%, and 7.2%, respectively), reduction in number of dead nodes (30.3%, 29.7%, 28.9%, 24.3%, 18%, and 11.5%, respectively), and network throughput (36.4%, 43.9%, 34.2%, 25%, 20%, 14.4%, respectively). Moreover, a supplementary experiment was conducted to test the efficiency of the HGORSA algorithm in dense and sparse networks, where the number of SNs was set at 50 and 500. The algorithm was evaluated based on the five standard aforementioned performance metrics. Furthermore, the robustness of HGORSA was validated using statistical measures, including standard deviation (Std), average (Avg), worst and best values, and box plots of the fitness function across 20 independent runs. Based on statistical results, HGORSA outperformed the other comparative meta-heuristics.