A New Approach To Efficiency Measurement: Hybrid Jaya Algorithm and Data Envelopment Analysis

Abstract

Efficiency measurement using linear programming is a crucial decision-making problem in operational research. Assessing efficiency scores via Stochastic Frontier Analysis (SFA), a parametric method, becomes challenging when decision-making units (DMUs) have multiple outputs, limiting the information's comprehensiveness. To address this issue, we propose an innovative alternative method that estimates SFA for DMUs with multiple outputs by defining multiple outputs as single outputs that yield the same efficiency. Combining DEA with heuristic methods like the JAYA algorithm offers a novel approach for converting multiple outputs into a single output when inputs and efficiency scores are provided. The JAYA algorithm used in this study includes two stopping criteria: convergence to a specified correlation value and reaching the maximum number of iterations. The proposed method successfully calculates the efficiency of DMUs for datasets of 20, 50, 97, and 100 units with correlation values of 95% and 99%. The JAYA algorithm computed 20, 50, 97, and 100 DMU samples in approximately 0.04, 0.45, 3.96, and 3.22 h, respectively (target correlation values of 95%). This method facilitates faster and more accurate efficiency measurements and enhances the ability to handle datasets with multiple outputs, providing a more robust and comprehensive analysis in operational research.