A quality of service distributed optimizer for Cognitive Radio Sensor Networks

详细信息    查看全文

• 作者：Tarek M. Salem^a ; ^{tareksalem@eri.sci.eg" class="auth_mail" title="E-mail the corresponding author} ; Salah Abdel-Mageid^b ; ^{smageid@azhar.edu.eg" class="auth_mail" title="E-mail the corresponding author} ; Sherine M. Abd El-kader^a ; ^{sherine@eri.sci.eg" class="auth_mail" title="E-mail the corresponding author} ; Mohamed Zaki^b ; ^{azhar@eun.eg" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Cognitive Radio Sensor Networks ; Multi-objective optimization ; Genetic Algorithms ; Non-dominated Sorting Genetic Algorithm ; Simulating Annealing
• 刊名：Pervasive and Mobile Computing
• 出版年：2015
• 出版时间：September 2015
• 年：2015
• 卷：22
• 期：Complete
• 页码：71-89
• 全文大小：2814 K

文摘

In Cognitive Radio Sensor Networks (CRSNs), a sensor node is provided with a cognitive radio unit to overcome the problem of frequency spectrum being crowded. Sensor nodes sense frequency gaps for Primary Users (PUs) to work as Secondary Users (SUs). However, Quality of Service (QoS) requirements for sensor nodes such as maximizing throughput and minimizing transmission power conflicts with minimizing interference between sensor nodes and PUs. Existing works have optimized QoS parameters considering frequency interference problem using Genetic Algorithms (GA) and Simulating Annealing (SA). In this paper, a distributed optimizer for CRSNs based on advanced multi-objective evolutionary algorithms named Non-dominated Sorting Genetic Algorithm (NSGA-II) has been proposed. A set of accurate fitness functions for NSGA-II implementation that fully control evolution of the algorithm have been employed. To the best of our knowledge, there is no published research in CRSN that contains all these intrinsic fitness functions in one system model. Simulation results show that the proposed optimizer can work as a distributed solution for CRSNs because it achieves a minimum number of iterations and minimum coverage time to reach an optimal solution compared to GA and SA. Such minimization matches the energy requirement for the underlying sensor nodes.