Data envelopment analysis with two distinct objectives of inputs or outputs

When data envelopment analysis (DEA) methodology is applied to benchmark selection, the outputs are frequently considered as derived from two distinct objectives. If the outputs (or inputs) derived from two distinct objectives are directly mixed together for executing DEA, decision making units (DMUs) with low outputs (or high inputs) of either objective may be identified as benchmark DMUs. Specifically, performance of benchmark DMUs may be dominated by that of the corresponding inefficient DMUs under either objective. Emulating such benchmarks to improve performance is problematic. Without involving value judgments or a prior information, this study develops a novel method, which we call Two-Objective DEA (TODEA), to identify the most appropriate benchmark DMUs whose performance is not dominated by the corresponding inefficient DMUs under either objective. To clarify the benefits of the developed method, the TODEA method is compared with a recently proposed benchmarking method, Two-Model DEA approach, via a numerical example.