Using fuzzy cognitive maps to identify multiple causes in troubleshooting systems

  • Authors:

  • Karl Perusich

  • Affiliations:

  • Purdue University, 1733 Mishawaka Ave., South Bend, IN 46634-7111, USA. Tel.: +1 574 520 5508/ Fax: +1 574 520 4286/ E-mail: Perusich@purdue.edu

  • Venue:
  • Integrated Computer-Aided Engineering
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Fuzzy cognitive maps are a qualitative tool that can capture the extensive cause/effect relationships that an expert believes exist within a complicated system such as an electronic circuit. In addition to predictive capabilities when inputs are applied and propagated through the model, the topology of the map itself can be used in diagnosing failures by identifying causes for nodes of interest. But simply tagging a node as a cause of another is inadequate because this identification does not indicate whether this node is a sole cause or whether other nodes must also be present. The identification of combinations of nodes is key to the determination of a strategy for locating the system problem. An algorithm is given in this paper that uses a modified version of the reachability matrix to identify multi-node combinations that can cause a node of interest.