ACCELERATING THE IDENTIFICATION OF EFFICIENT FRONTIER FOR LARGE-SCALE DATA ENVELOPMENT ANALYSIS

Data Envelopment Analysis (DEA) is a linear programming (LP)-based technique used to evaluate the performance of homogeneous Decision-Making Units (DMUs) with multiple inputs and outputs. However, applying DEA to large-scale datasets with a high number of DMUs poses significant computational challenges. A practical solution to this problem is to solve the LPs of DEA with only decision variables associated with efficient DMUs. Building on this idea, we propose an arithmetic approach to identify a subset of efficient DMUs. Using this subset, we can effectively classify DMUs and rapidly identify the complete set of efficient units. To validate our approach, we conducted experiments on both real-world and simulated datasets, varying the cardinalities (number of DMUs), dimensions (number of inputs and outputs), and densities (proportion of efficient DMUs). The results demonstrate the effectiveness and robustness of our method, highlighting its potential for improving existing DEA methodologies.