Fuzzy cognitive network: A general framework

  • Authors:

  • Theodoros L. Kottas;Yiannis S. Boutalis;Manolis A. Christodoulou

  • Affiliations:

  • Automatic Control Systems Laboratory, Department of Electrical & Computer Engineering, Democritus University of Thrace, 67100 Xanthi, Greece;Automatic Control Systems Laboratory, Department of Electrical & Computer Engineering, Democritus University of Thrace, 67100 Xanthi, Greece;Department of Electronic and Computer Engineering, Technical University of Crete,73100, Chania, Crete, Greece

  • Venue:
  • Intelligent Decision Technologies
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we present a general computational and operational framework for the Fuzzy Cognitive Network FCN, which is a direct extension of Fuzzy Cognitive Maps FCM. The proposed framework assumes a network operation, which continuously receives feedback from the system it describes and outputs control or decision values. This way, its knowledge is continuously updated making it suitable for adaptive decision making or even for adaptive control tasks. The interconnection weights are continuously updated based on a modified delta rule that provides smooth and fast convergence and prevents the concept and weight values from being saturated. To avoid intensive interference of the updating mechanism with the real system, a technique is proposed that stores the previously encountered operational situations in a fuzzy rule database. The explanation of the proposed methodology is interweaved with the FCN description of a simulated hydro-electric plant, which is also used for the experimental results. The proposed framework can be used both for on-line control and decision making tasks.