Supplementary Sequential 3WD-EDAS-E.pdf

Last-mile delivery (LMD) has become a critical focus in supply chain management, primarily due to multiple factors influencing its operational efficiency. Evaluating its performance presents a multi-criteria decision-making challenge, requiring systematic approaches to ensure informed decision-making. This study contributes to the evolving literature on LMD evaluation by introducing methodological improvements within the EDAS (Evaluation based on Distance from the Average Solution) method to generate more intuitive insights into decision-making. First, instead of the L1-metric used in the canonical EDAS, we introduced Euclidean-based EDAS (EDAS-E) to better capture the underlying geometry of the space generated by a relatively high number of decision criteria in LMD assessments. Second, adopting the notion of "thinking in threes," we developed an algorithm to implement a sequential three-way decision (3WD) that enhances explainability and addresses ambiguity in the evaluation process. A study on LMD, lifted from the literature, is used to demonstrate the efficacy of the proposed sequential 3WD-EDAS-E, guiding decision-makers in systematically distinguishing between acceptable, unacceptable, and indeterminate alternatives. The comparative analysis shows high agreement between the results generated by the proposed approach and those obtained from other multi-criteria methods, reinforcing its robustness and applicability.

最后一公里配送（Last-mile delivery, LMD）已成为供应链管理中的核心关注点，这主要源于诸多影响其运营效率的因素。对其绩效开展评估属于多准则决策难题，需借助系统性方法以保障决策的科学性与合理性。本研究针对持续发展的LMD评估相关文献做出贡献：通过对基于距离平均解的评估法（Evaluation based on Distance from the Average Solution, EDAS）进行方法学改进，以生成更具直观性的决策洞察。其一，摒弃经典EDAS所采用的L1范数，我们引入了基于欧氏距离的EDAS（EDAS-E），以更好地捕捉LMD评估中由相对较多决策准则所生成空间的内在几何特性；其二，借鉴“三分思考”的理念，我们开发了一种实现序贯三支决策（3WD）的算法，该算法可提升评估过程的可解释性并解决其中的模糊性问题。本研究采用一篇已发表文献中的LMD案例，用以验证所提出的序贯三支决策-EDAS-E方法的有效性，辅助决策者系统性地区分可接受、不可接受及待定三类备选方案。对比分析结果显示，所提方法生成的结果与其他多准则方法所得结果具有高度一致性，印证了该方法的稳健性与适用性。