LAUSR

Data envelopment analysis (DEA) method has been widely used in many economic and industrial applications to measure efficiency and rank performances of decision making units (DMUs). Improving the accuracy and computation time in measuring the efficiency of DMUs have been two main challenges for the DEA. Specifically, with large DMUs, the DEA-based methods are argued to require large amount of memory space and CPU time to measure DMUs efficiencies, and suffer from inability to obtain complete performance ranking. To address these issues, in this paper, a new alternative method that is based on input oriented model (IOM) and efficiency ratio (ER), called ratio efficiency dominance (RED), is proposed. The proposed method seeks to minimize the inputs while maximizing the outputs to obtain efficiency or performance scores, which is independent of DEA method and the use of linear programming (LP). It is also to overcome the drawbacks of uncontrolled convergence, non-generalization and instability induced from integrating prediction techniques such as neural networks (NNs) with DEA. To evaluate the proposed method, experiments were performed on small, large and very large DMUs data sets to show the effectiveness of proposed method. The experimental results demonstrated that, in all cases, the proposed method is able to produce a complete and more accurate ranking compared to the conventional DEA methods or its hybrids.