This paper is concerned with how to stably arrange the total input resources and output productions when the total outputs can be forecasted in a centralized decision-making environment based on data envelopment analysis (DEA). It usually involves the participation of all individual decision-making units, each contributing in part to the total production. Two stability indexes are defined to assess the stability of production planning, i.e. scale stability, efficiency and technology stability. Then, we propose a DEA-based model to derive the next period production planning. Finally, we illustrate the proposed approach and compare the results between two existing studies and this paper by using three numerical examples. The results show that the proposed production planning in this paper is stabler than existing studies. It implies that the production planning in this paper would be more practicable for real production.

本文涉及在基于数据包络分析（DEA）的集中决策环境中预测总产出时如何稳定地安排总投入资源和产出。它通常涉及所有单个决策部门的参与，每个决策部门都对总产量做出部分贡献。定义了两个稳定性指标来评估生产计划的稳定性，即规模稳定性，效率和技术稳定性。然后，我们提出了一个基于DEA的模型，以得出下一阶段的生产计划。最后，我们使用三个数值示例说明了所提出的方法，并比较了两个现有研究与本文的结果。结果表明，本文提出的生产计划比现有研究更稳定。