یک پروتکل مسیریابی کارآمد انرژی مبتنی بر سیستم فازی ژنتیکی بر اساس منطقه بهینه سازی شده برای شبکه های حسگر موبایل (JEEP)

Wireless sensor networks have become increasingly popular because of their ability to cater to multifaceted applications without much human intervention. However, because of their distributed deployment, these networks face certain challenges, namely, network coverage, continuous connectivity and bandwidth utilization. All of these correlated issues impact the network performance because they define the energy consumption model of the network and have therefore become a crucial subject of study. Well-managed energy usage of nodes can lead to an extended network lifetime. One way to achieve this is through clustering. Clustering of nodes minimizes the amount of data transmission, routing delay and redundant data in the network, thereby conserving network energy. In addition to these advantages, clustering also makes the network scalable for real world applications. However, clustering algorithms require careful planning and design so that balanced and uniformly distributed clusters are created in a way that the network lifetime is enhanced. In this work, we extend our previous algorithm, titled the zone-based energy efficient routing protocol for mobile sensor networks (ZEEP). The algorithm we propose optimizes the clustering and cluster head selection of ZEEP by using a genetic fuzzy system. The two-step clustering process of our algorithm uses a fuzzy inference system in the first step to select optimal nodes that can be a cluster head based on parameters such as energy, distance, density and mobility. In the second step, we use a genetic algorithm to make a final choice of cluster heads from the nominated candidates proposed by the fuzzy system so that the optimal solution generated is a uniformly distributed balanced set of clusters that aim at an enhanced network lifetime. We also study the impact and dominance of mobility with regard to the variables. However, before we arrived at a GFS-based solution, we also studied fuzzy-based clustering using different membership functions, and we present our understanding on the same. Simulations were carried out in MATLAB and ns2. The results obtained are compared with ZEEP.