Increasing energy efficiency of rule-based fuzzy clustering algorithms using CLONALG-M for wireless sensor networks

Sert S. A., Yazici A.

Applied Soft Computing, vol.109, 2021 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 109
  • Publication Date: 2021
  • Doi Number: 10.1016/j.asoc.2021.107510
  • Journal Name: Applied Soft Computing
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Applied Science & Technology Source, Compendex, Computer & Applied Sciences, INSPEC
  • Keywords: Clonal selection principle, Performance tuning, Fuzzy function approximation, Fuzzy clustering algorithms, Wireless sensor networks, MEMBERSHIP FUNCTIONS, SELECTION, OPTIMIZATION
  • Middle East Technical University Affiliated: Yes


© 2021 Elsevier B.V.Because of its efficiency, clustering is used for effective communication in Wireless Sensor Networks (WSNs). In the WSN clustering area, fuzzy approaches are found to be superior to crisp cluster counterparts when the boundaries between clusters are unclear. As a result, many studies have proposed some fuzzy-based solutions to the cluster problem in WSNs. Most rule-based fuzzy clustering systems employ field experts in trial and error processes, identifying and defining fuzzy rules as well as the forms of membership functions at the output; thus, considerable time has been allocated to realize and define these functions. Therefore, it is almost impossible or impractical to achieve a fuzzy system optimally. In this study, we propose a modified clonal selection algorithm (CLONALG-M) to improve the energy efficiency of rule-based fuzzy clustering algorithms. Although some studies in the literature focus on fuzzy optimization in general, to the best of our knowledge, performance improvement of rule-based fuzzy clustering algorithms is not taken into account. The CLONALG-M algorithm based on the Clonal Selection Principle is used to elucidate the basic principles of an adaptive immune system. In this study, we apply this principle to determine the approximate deployment of output-based membership functions that increase the performance of rule-based fuzzy clustering algorithms, whose rule base and shape of membership functions are previously known. Experimental analysis and evaluations of the proposed approach have been performed on selected fuzzy clustering approaches, and obtained results show that our approach performs and adapts well for improving performance of fuzzy output functions.