面向产品质量设计的语言群决策方法研究：基于群体异质性视角

Product quality design is to identify quality requirements in the product design, and to define product quality features that could meet customer needs. In order to explore potential customer needs and design innovative products, enterprises need to collect multi-sourced heterogeneous data of customer needs and analyze them in depth. Besides, multiple stakeholders have to take part in the vital activities including competitive analysis, risk analysis, critical purchase decisions and so on, for reducing the decision failure risk. With respect to the aforementioned issues, information providers could depict their judgments in a convenient and flexible way by using linguistic evaluation values. Therefore, this project is planning to study linguistic group decision-making problems in the stage of product quality design from the perspective of group heterogeneity. Firstly, this project intends to collect linguistic evaluation data from customers by crowdsourcing, and establish the cloud-model based information integration approach to convert heterogeneous linguistic data into probabilistic linguistic information. Then, the factor analysis approaches and multi-criteria decision-making approaches based on probabilistic linguistic information, will be proposed to deal with the corresponding decision problems in the stage of product quality design. Finally, forest food will be chosen in empirical study for verifying the proposed models and approaches, because this product is critical for targeted poverty alleviation. The research outcomes of this project could enrich the linguistic group decision theory, and improve decision quality in the stage of product quality design.

产品质量设计是指在产品设计中提出质量要求，确定满足顾客需求的产品质量特征。为了发掘顾客潜在的需求、设计创新性产品，企业需要收集并深入分析多源异构的顾客需求数据。同时，产品质量设计阶段的竞争分析、风险分析、关键采购决策等，常常需要多个利益相关者共同参与，以降低决策失败的风险。在上述群体决策问题中，语言评价值能让信息提供者以灵活便捷的方式描述自己的判断，因此本项目将充分考虑群体的异质性特点，提出面向产品质量设计的语言群决策方法。首先，通过众包模式采集顾客的语言评价数据，利用基于云模型的异质语言信息融合方法得到概率语言信息。然后，提出基于概率语言信息的因素分析方法和多准则决策方法，以解决产品质量设计阶段相应的决策问题。最后，选择有助于精准扶贫的森林食品，对提出的模型及方法进行实证研究。本项目研究成果不仅能丰富语言群决策理论，还能提高产品质量设计阶段的决策质量。

产品质量设计是指在产品设计中提出质量要求，确定满足顾客需求的产品质量特征。为了发掘顾客潜在的需求、设计创新性产品，企业需要收集并深入分析多源异构的顾客需求数据。同时，产品质量设计阶段的竞争分析、风险分析、关键采购决策等，常常需要多个利益相关者共同参与，以降低决策失败的风险。在上述群体决策问题中，语言评价值能让信息提供者以灵活便捷的方式描述自己的判断，但在数据融合时需要考虑群体的异质性特点。由此，本项目针对产品质量设计阶段的语言群决策方法展开研究，应用概率语言决策理论与方法解决相关问题，主要内容包括：提出基于云模型的异质语言信息融合方法，将采集的群体语言评价数据转化为概率语言信息；提出基于概率语言信息的因素分析方法和多准则决策方法，以解决产品质量设计阶段相应的决策问题；选择典型的森林食品开展新模型和新方法的实证研究，对其进行验证和完善。本项目研究成果能够丰富语言群决策理论，辅助提升产品质量设计阶段的决策质量，促使最终的产品质量更加贴近顾客的实际需求。