A New Method for Improving the Discrimination Power and Weights Dispersion in the Data Envelopment Analysis

The appropriate choice of input-output weights is necessaryto have a successful dea model. generally, if the number ofdmus i.e., n, is less than number of inputs and outputs i.e., m+s,then many of dmus are introduced as efficient then the discriminationbetween dmus is not possible. besides, dea models are free tochoose the best weights. for resolving the problems that are resultedfrom freedom of weights, some constraints are set on the input-outputweights. symmetric weight constraints are a kind of weight constrains.in this paper, we represent a new model based on a multi-criteriondata envelopment analysis (mcdea) are developed to moderate thehomogeneity of weights distribution by using symmetric weight constrains.consequently, we show that the improvement of the dispersalof unrealistic input-output weights and the increasing discriminationpower for our suggested models. finally, as an application of the newmodel, we use this model to evaluate and ranking guilan selected hospitals.