Data envelopment analysis (DEA) relies on two main postulates of convexity and inefficiency (free disposability). No free disposability postulate is suggested to address undesirable measures. In this study, we demonstrate how no-disposability assumption can be correctly integrated into the DEA framework. We propose the appropriate constraints that should be used in the absence of the free disposability postulate in a DEA model. The additional constraints bound the previously unbounded feasible region (production technology) rather than altering the strongly efficient frontier. We also discuss that treating an undesirable output (input) as a desirable input (output) does not affect the corresponding efficient frontier of a dataset, but misrepresents its corresponding production technology in the presence of free disposability postulate. We provide numerical examples to clarify the concerns in treating an undesirable measure as a desirable measure. A real-life example of United States’ electric power plants is also discussed.

中文翻译：

---

在数据包络分析中实现无自由可处置性


数据包络分析 （DEA） 依赖于凸性和低效率（自由可处置性）的两个主要假设。不建议使用免费可丢弃性假设来解决不良措施。在这项研究中，我们展示了如何将非一次性假设正确集成到 DEA 框架中。我们提出了在 DEA 模型中不存在自由可处置性假设的情况下应该使用的适当约束。额外的约束限制了以前无界的可行区域（生产技术），而不是改变强有效边界。我们还讨论了将不需要的输出（输入）视为理想的输入（输出）不会影响数据集的相应有效边界，但在存在自由可抛弃性假设的情况下歪曲了其相应的生产技术。我们提供了数值示例，以阐明将不良度量视为理想度量的担忧。还讨论了美国发电厂的一个真实示例。