摘要
Assessing the efficiency of decision-making units (DMU) with fixed-sum outputs in a two-stage parallel-series structure is challenging for decision-makers. The two main challenges are: (1) constructing an equilibrium efficient frontier (EEF) in a two-stage parallel-series structure and, (2) handling multiple feasible EEFs. This study embeds the stochastic multi-criteria acceptability analysis (SMAA) method into a data envelopment analysis (DEA) model framework and proposes a two-stage parallel-series SMAA-DEA approach. This approach explores feasible EEFs and proposes decision-making indicators to obtain rank acceptability indices, efficiency dominance probability, and holistic rank acceptability for DMUs. Our approach's effectiveness and rationality are illustrated using a real dataset of 24 non-life insurance companies. The proposed approach provides reliable ranking results for DMUs and rich information to assist decision-making.