Training Fuzzy Cognitive Maps via Extended Great Deluge Algorithm with applications

  • Authors:

  • Adil Baykasoglu;Zeynep D. U. Durmusoglu;Vahit Kaplanoglu

  • Affiliations:

  • University of Gaziantep, Faculty of Engineering, Department of Industrial Engineering, Gaziantep, Turkey;University of Gaziantep, Faculty of Engineering, Department of Industrial Engineering, Gaziantep, Turkey;University of Gaziantep, Faculty of Engineering, Department of Industrial Engineering, Gaziantep, Turkey

  • Venue:
  • Computers in Industry
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Fuzzy Cognitive Map (FCM) is an extension of classical cognitive map (CM). It is mainly a soft computing technique which is used to represent knowledge and causal inference. In order to develop a FCM for a system, a group of experts are usually asked to define concepts or factors that represent the system and describe relations among these concepts. However, in many cases FCM can include subjective factors involved in the determination of FCM weights. Several training (or learning) algorithms are employed in the literature to reduce the subjectivity of the inference so far. In this study, Extended Great Deluge Algorithm (EGDA) has been considered first time in the literature as a training algorithm for FMCs. The performance of the algorithm has been tested with two problems. The first problem is selected from the literature which is a ''industrial process control problem''. For this problem the proposed algorithm provided promising results. In the second problem a simulation model of a job shop is developed and utilized in order to investigate causal relationship between the control/performance factors through FCM.