Abstract: Efficiency measurement in traditional Data Envelopment Analysis (DEA) models depends mainly on sampling, and the reliability of efficiency results is directly affected by the data quality of the Decision-making Units (DMUs). The interval DEA models have been developed to address the problem of inaccurate data in DEA. However, this study reveals that the existing models and approaches may lead to inaccuracies and inconsistencies in the measurement of efficiency and congestion effect. To address these shortcomings, this paper first proposes a new interval grey number envelopment model which takes into account the cross-evaluation logic between DMUs in an uncertain environment. It then proposes a novel algorithm for identifying and measuring the effects of congestion. The new method reserves the maximum possible range of efficiency values and improves the congestion identification between DMUs. Finally, the advantages and improvements of the proposed method are illustrated by several numerical examples.

Key words: Interval data envelopment analysis, Interval grey number, Efficiency measurement, Congestion identification