This study evaluated the benchmarking process of active queue management (AQM) methods, which consider a multicriteria decision‐making (MCDM) problem using multidimensional criteria. Academic studies have benchmarked the AQM methods using MCDM techniques. However, these studies have used existing MCDM techniques, which face considerable theoretical challenges. The latest MCDM method called fuzzy decision by opinion score (FDOSM) was published in the Journal of Applied Soft Computing in 2020 to address the theoretical challenges of the existing MCDM methods. However, FDOSM continues to encounter serious issues. That is, it exclusively depends on the direct aggregation MCDM approach based on arithmetic mean (AM) operator. However, performing other operators (i.e., geometric mean, harmonic mean, and root mean square), in addition to applying other MCDM approaches (i.e., distance measurement and compromise rank), may result in different ranking results. Hence, this study mainly proposes an extension of FDOSM through the following aspects: (1) application of different aggregation techniques in the direct aggregation MCDM approach, (2) discussion of the effectiveness of each type on the final AQM benchmarking, and (3) use of varying MCDM approaches on FDOSM to reach the optimum result when benchmarking the AQM methods. The current research methodology is based on two sequential phases. The first phase provides the decision matrix used in benchmarking the AQM methods. The decision matrix was constructed based on the AQM evaluation criteria and a list of AQM methods. The second phase presents two stages, namely, data transformation unit and data processing. Findings of the AQM benchmarking are as follows. (1) In the individual FDOSM, two main configurations are recommended when using the AQM benchmarking: direct aggregation MCDM approach with AM operator and compromise rank approach. Benchmarking results of both configurations based on six decision makers are nearly similar, with the AQM BLUE method being ranked the best. The exception is for the results of the compromise rank approach based on the third decision maker, which revealed that the AQM ERED method is the best. (2) Results of the group FDOSM showed a relatively similar order for the AQM methods in both configurations, with the AQM BLUE method being the best. (3) Lastly, significant differences were found among the groups' scores, thereby indicating the validity of the FDOSM‐based AQM benchmarking results.

中文翻译：

---

基于意见评分法扩展模糊决策的网络拥塞控制主动队列管理方法多维标杆

本研究评估了主动队列管理 (AQM) 方法的基准测试过程，该方法使用多维标准考虑多标准决策 (MCDM) 问题。学术研究已经使用 MCDM 技术对 AQM 方法进行了基准测试。然而，这些研究使用了现有的 MCDM 技术，面临着相当大的理论挑战。最新的 MCDM 方法称为基于意见评分的模糊决策 (FDOSM)，于 2020 年发表在《应用软计算杂志》上，以解决现有 MCDM 方法的理论挑战。但是，FDOSM 继续遇到严重问题。也就是说，它完全依赖于基于算术平均 (AM) 算子的直接聚合 MCDM 方法。但是，执行其他运算符（即几何平均数、调和平均数和均方根），除了应用其他 MCDM 方法（即距离测量和折衷排名）外，可能会导致不同的排名结果。因此，本研究主要通过以下几个方面提出对 FDOSM 的扩展：（1）不同聚合技术在直接聚合 MCDM 方法中的应用，（2）讨论每种类型对最终 AQM 基准测试的有效性，以及（3）在对 AQM 方法进行基准测试时，在 FDOSM 上使用不同的 MCDM 方法以达到最佳结果。当前的研究方法基于两个连续的阶段。第一阶段提供用于对 AQM 方法进行基准测试的决策矩阵。决策矩阵是基于 AQM 评估标准和一系列 AQM 方法构建的。第二阶段分为数据转换单元和数据处理两个阶段。AQM 基准测试的结果如下。(1) 在单个 FDOSM 中，使用 AQM 基准测试时推荐两种主要配置：使用 AM 运营商的直接聚合 MCDM 方法和折衷秩方法。基于六位决策者的两种配置的基准测试结果几乎相似，其中 AQM BLUE 方法排名最佳。例外是基于第三个决策者的折衷等级方法的结果，它表明 AQM ERED 方法是最好的。(2) FDOSM 组的结果显示两种配置中 AQM 方法的顺序相对相似，其中 AQM BLUE 方法最好。(3) 最后，各组得分之间存在显着差异，从而表明基于 FDOSM 的 AQM 基准测试结果的有效性。