IMPROVING THE SELECTION SYMMETRIC WEIGHTS AS A SECONDARY GOAL IN DEA CROSS-EFFICIENCY EVALUATION

Recently, some authors proposed the use of symmetric weightsfor computing the elements of cross-efficiency matrix. In spite ofthe fact that the proposed method decreases the number of zeroweights, a large number of zero weights may still exist among inputand output symmetric weights. To decrease the number of input andoutput symmetric weights, this paper improves the proposed secondarygoal model. The improved method generates more acceptable results inthe ranking process of the decision making units (DMUs). Finally, anumerical example is provided to illustrate the applicability of theproposed method.