A New Way to Determine Common Set of Weights for Full Rank of Performance of Decision Making Units


ALP İ.

7th International Eurasian Conference on Mathematical Sciences and Applications (IECMSA), Kyiv, Ukraine, 28 - 31 August 2018, vol.2037 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 2037
  • Doi Number: 10.1063/1.5078456
  • City: Kyiv
  • Country: Ukraine

Abstract

Traditional data envelopment analysis (DEA) models select weights specific for every decision making unit (DMU) in a way that maximize the performance of each DMU. With the DEA models, the inputs and outputs of each DMU are evaluated with the different set of weights that are not common. Importances of weights of the inputs and outputs not to happen same for every DMU. This is advantageous for some DMUs, while for other DMUs it is disadvantageous. Another drawback is that in the DEA performance calculations, for some inputs and outputs, it selects very small or zero weights. A very small near zero or zero weight probably means that an important criterion will not be considered in the performance calculation. Together with above, another defect is the same efficiency score (1/100) are given to all efficient DMUs. This prevents full ranking of DMUs. One way for eliminate the disadvantages which mentioned above is to use same set of weights during calculation of the performance of all DMUs. The weights of the Andersen -Petersen super efficiency model were used as stepping stone in this new common set of weights (CSWs) generation algorithm. This new algorithm will be apply to the well-known data of a real -world problem in litarature.