DEA-based robot selection procedure incorporating fuzzy criteria values


KARSAK E. E.

1999 IEEE International Conference on Systems, Man, and Cybernetics 'Human Communication and Cybernetics', Tokyo, Japan, 12 - 15 October 1999, vol.1 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 1
  • City: Tokyo
  • Country: Japan
  • Galatasaray University Affiliated: Yes

Abstract

Data Envelopment Analysis (DEA), which avoids the critical assumption that the performance parameters are mutually independent, appears as an effective decision tool for the robot selection problem. However, a robust robot selection procedure necessitates the consideration of both quantitative criteria such as cost and engineering attributes, and qualitative criteria, e.g. vendor-related attributes, in the decision process. Qualitative attributes may be indicated by using linguistic variables or fuzzy numbers. This paper introduces a revised DEA formulation incorporating both quantitative and qualitative factors to determine the best robot alternative. When multiple efficient robot alternatives are identified as a result of the analysis, cross efficiency analysis is used to discriminate between the alternatives, and thus, to determine the best robot alternative. A comprehensive example is presented to illustrate the decision-making procedure.