In this paper, we present a recurrent neural network model for solving CCR Model in Data Envelopment Analysis (DEA). The proposed neural network model is derived from an unconstrained minimization problem. In the theoretical aspect, it is shown that the proposed neural More
In this paper, we present a recurrent neural network model for solving CCR Model in Data Envelopment Analysis (DEA). The proposed neural network model is derived from an unconstrained minimization problem. In the theoretical aspect, it is shown that the proposed neural network is stable in the sense of Lyapunov and globally convergent to the optimal solution of CCR model. The proposed model has a single-layer structure. A numerical example shows that the proposed model is effective to solve CCR model in DEA.
Manuscript profile
Data envelopment analysis (DEA) is a non-parametric programming method for evaluating the relative efficiency of a set of peer decision-making units (DMUs) with multiple inputs and multiple outputs. The DEA cross-efficiency method is a well-known method that use to eval More
Data envelopment analysis (DEA) is a non-parametric programming method for evaluating the relative efficiency of a set of peer decision-making units (DMUs) with multiple inputs and multiple outputs. The DEA cross-efficiency method is a well-known method that use to evaluate and ranking a set of peer decision-making units. Whenever a DMU intends to evaluate other DMUs, it faces the problem of non-uniqueness optimal weights of DEA models. Because different weights give us different cross-scores and subsequently different cross-efficiencies scores and this will confuse the decision-maker to make an ultimate decision. The main drawback of this method is the alternate optimal solution set of the DEA model. The main purpose of this study is to propose an approach to this problem to generate non-dominated DEA cross-efficiency scores. We propose a linear programming secondary goal model to select a set of optimal weights for each DMU. Our proposed method is not only simpler than other methods presented with the same purpose, but also does not go beyond the main method.
Manuscript profile
