Abstract:

In case of multiple inputs and outputs, performance of Decision Making Units (DMU) is defined as the ratio of weighted sum of outputs to weighted sum of inputs. There are two group ways to determine the weights of performance : objective and subjective approaches mainly. In the subjective approaches, weights which will be given to the inputs and outputs are determined based on the opinion of DMUs or experts. In the objective approaches, weights are found via models  and calculations which are not based on personal judgments. One of them is the most important and widely used Data Envelopment Analysis (DEA) method. Data Envelopment analysis is a nonparametric and operations research-based technique. DEA, in the performance calculations, assigns weights to multiple inputs and outputs in an objective manner by means of a linear programming model to maximize the performance of each DMU. There may be two disadvantages for the weights which calculated by this method: I.  To give very small or zero weights to important inputs and outputs. II.  In aggregate evaluation, computed weights generally to be different for each input and output for different decision- makers; in the performance evaluation, importances or weights of the inputs and outputs not to happen same for every DMU. One way for eliminate the disadvantages mentioned above is to use common weights when calculating the performance of DMUs. Another method is to use the correlation coefficients between inputs and outputs. Mentioned methods in this work will be interpreted by applying to the data of a real-world problem.

Keywords: