Node centrality and average weight.

In intelligent manufacturing for complex products, the configuration and allocation of human-robot collaboration units (HRCUs) are of critical importance for enhancing production performance. To address the insufficient research on the impact of individual irrational behaviors and group-reference behaviors in HRCUs construction, a stable one-to-many human-robot-position matching decision-making (HRPMDM) method in hesitant fuzzy environments is proposed. Specifically, linguistic hesitant fuzzy sets (LHFSs) are adopted to characterize evaluators’ dual uncertainties in linguistic term selection and membership degree assignment. Subsequently, the Cloud Model is adopted to quantitatively transform the LHFSs, thereby providing support for the proposed clustering algorithm based on cognitive similarity, enabling it to divide the matching objects into several subgroups according to the degree of cognitive similarity among individuals. Furthermore, to reduce the bias in attribute weight assessment caused by peer effects, a social network-based algorithm that enables precise quantification of subgroup and member weights is proposed. These quantified weights are then integrated into the group consensus adjustment process to provide reliable reference correction values for individual assessments. Additionally, multi-proposition belief structures are introduced to represent uncertain matching preference rankings (UMPRs) influenced by group reference behaviors, and a corresponding satisfaction measurement method is further developed. Finally, a practical case study demonstrates the operational feasibility and performance efficacy of the proposed method. This study is the first to integrate carbon neutrality cost optimization objectives into human-robot matching decisions and develops a Quality of Service (QoS)-optimized allocation strategy for HRCUs in heterogeneous production environments. The results demonstrate that the proposed matching method has led to significant improvements in both production efficiency and environmental sustainability for complex product manufacturing.

在复杂产品智能制造领域，人机协作单元（Human-Robot Collaboration Units, HRCUs）的配置与分配，对提升生产绩效具有关键意义。针对当前人机协作单元构建过程中，个体非理性行为与群体参照行为的影响研究尚存不足的问题，本文提出一种犹豫模糊环境下的稳定一对多人机-岗位匹配决策（Human-Robot-Position Matching Decision-Making, HRPMDM）方法。具体而言，研究采用语言犹豫模糊集（Linguistic Hesitant Fuzzy Sets, LHFSs），刻画评估者在语言术语选择与隶属度赋值环节的双重不确定性。随后，借助云模型对语言犹豫模糊集进行定量转换，为所提出的基于认知相似度的聚类算法提供支撑，使该算法可依据个体间的认知相似度，将匹配对象划分为若干子群体。此外，为削弱同伴效应引发的属性权重评估偏差，本文提出一种基于社交网络的算法，可实现子群体与成员权重的精准量化。随后将这些量化后的权重融入群体共识调整流程，为个体评估提供可靠的参考修正值。另外，引入多命题置信结构，表征受群体参照行为影响的不确定匹配偏好排序（Uncertain Matching Preference Rankings, UMPRs），并进一步构建了对应的满意度测度方法。最后，通过实际案例验证了所提方法的运行可行性与性能有效性。本研究首次将碳中和成本优化目标纳入人机匹配决策范畴，并针对异构生产环境下的人机协作单元，构建了服务质量（Quality of Service, QoS）优化配置策略。研究结果表明，所提出的匹配方法可显著提升复杂产品制造的生产效率与环境可持续性。