Reduction Difference Between the Profile weights in Stochastic cross-efficiency

Cross-efficiency method is a useful tool for efficiency evaluation of decision-making units in data envelopment analysis. The issue of non-uniqueness of optimal weights in the cross-efficiency evaluation has reduced the usefulness of this powerful method. This paper introduces a new method for selection of weights profiles as the secondary goal in cross-efficiency with stochastic data. The issue of zero-weight which implies the exclusion of some variables from the assessments, is also addressed in the new proposed method. The provided weights selection method also reduces the weight disparity in the achieved weights profile. In the peer-restricted stochastic cross-efficiency evaluation, the new approach guarantees that different DMUs should not attach very different weights to the same variables. As the result, a common set of weights using the idea of similarity between sets of weights is achieved in the proposed computation method. Some numerical examples are also used for illustration and comparison purposes.