【作者】 义红云；

【导师】 徐晓燕；

【作者基本信息】 中国科学技术大学 ， 管理科学与工程， 2009， 硕士

【摘要】 抽样检验在生产的任何阶段,包括生产过程中以及过程末的检验,都对产品质量评估有非常重要的作用。之前的抽检方案设计方法一般是产品生产商与客户之间的非结构化谈判过程。在抽检设计过程中,生产者认为生产商风险α越小越好,而客户认为客户风险β越小越好,同时又必须满足双方的经济要求(利益最大或成本最小)。传统的抽样设计方案不能很好的协调双方的利益冲突。基于此问题,本文对此展开研究。首先,本文通过对抽样设计理论的回顾以对抽样设计的研究现状做一个较为全面的了解,并且研究了多目标抽样的指标筛选问题。其次研究了多目标抽样的样本容量问题,在考虑成本因素时给出了多目标抽样的最佳样本容量确定方法。最后引入检出期望和检出方差,从一个新的角度来评价抽样方案。然后,我们介绍了质量功能展开(Quality Function Deployment,简记QFD)的基本概念和原理以及三种重要的问题解决系统。并重点研究了多目标抽样设计情形下的质量屋,同时对质量屋中的顾客要求的确定和权重排序方法以及相关矩阵做了比较详细的研究。最后,文中对基于QFD的多目标抽样设计给出了计数和计量两种情形下的抽样设计方法分析,并给予了一个案例分析。文章的主要创新点有以下几点:首先,本文首次将QFD与多目标抽样相结合以设计一个更好的抽样方案。其次,之前的抽检方案设计方法一般是产品生产商与客户之间的非结构化谈判过程,本文基于QFD矩阵提出一种结构化设计方法,将产品的生产商和客户统一视作QFD矩阵的顾客,并将定性的顾客要求转化为定量的抽样质量特征,以满足顾客要求。最后,本文将QFD矩阵和合作博弈相结合以设计一个满意抽检方案,协调生产商和客户的利益冲突并满足其经济要求。更多还原

【Abstract】 Acceptance sampling procedures play a vital role in assessing the product quality at any stage of the manufacturing, both during in-process inspection and during final inspection of the products. Previous methods for designing a sampling plan are of unstructured negotiation process between the producer and the user of a product. During the sampling design process, the producer desires smallerα, the user wants smallerβ, and maximize the profit or minimize the cost. The traditional sampling plan can’t harmonize both sides’ benefit conflicts. Based this question, we will study this problem.Firstly, based on the review of sampling theory, we will get a full understanding of the research status about sampling theory, and studied how to select the indicators about multi-object sampling plan. Then we studied the problem about multi-object sampling plan sample size, and give a best method to determine the sample size under cost factor. Lastly introduce two important concepts: outgoing expectation and outgoing variance, to present a new appraisal index of adjust sampling plan.Then we introduced the basic concept and theory about quality function deployment (QFD), and also introduced three important question solving systems. Then focused on the house of quality (HOQ) under multi-object sampling plan and we study the problem how to determine the customer requirements, the relative weights and correlation matrix in house of quality.Lastly, this paper gives two cases: count-based and measurement-based multi-object sampling plan based QFD.The new ideas are as the following:Firstly, this paper combines the QFD and multi-object sampling plan to design a good sampling plan for the first time. Secondly, previous methods for designing a sampling plan are of unstructured negotiation process between the producer and the user of a product. By applying the QFD matrix, this paper proposes a structured sampling plan designing approach. In the proposed approach, both the producer and the user of a product are treated as the customers of a QFD matrix, and thus customer voices (qualitative) are translated into sampling plan specifications (quantitative) directed at customer satisfaction. Lastly, QFD and cooperative game are adopted to design a structured sampling plan to satisfy the customer’s intrinsic requirements and harmonize the benefit conflicts.更多还原