رویکرد برنامه ریزی هدف خطی برای تعیین وزن اهمیت نسبی نیاز مشتری در گسترش کارکرد کیفیت
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|7191||2011||11 صفحه PDF||سفارش دهید||7560 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Information Sciences, Volume 181, Issue 24, 15 December 2011, Pages 5523–5533
Quality function deployment (QFD) is a planning tool used in new product development and quality management. It aims at achieving maximum customer satisfaction by listening to the voice of customers. To implement QFD, customer requirements (CRs) should be identified and assessed first. The current paper proposes a linear goal programming (LGP) approach to assess the relative importance weights of CRs. The LGP approach enables customers to express their preferences on the relative importance weights of CRs in their preferred or familiar formats, which may differ from one customer to another but have no need to be transformed into the same format, thus avoiding information loss or distortion. A numerical example is tested with the LGP approach to demonstrate its validity, effectiveness and potential applications in QFD practice.
Quality function deployment (QFD) is a total quality management (TQM) tool that translates the voice of the customer into the voice of the engineer. It has been successfully applied in various industries. To implement QFD, customer requirements (CRs) have to be identified and assessed. In particular, determining a set of correct importance weights for CRs is important because the importance weights have significant effects on the target values of engineering characteristics. Various methods have been suggested in QFD literature to determine the relative importance weights of CRs. The simplest method to prioritize CRs is the use of the L-point Likert scale, e.g., 1–5, however, this method is considered unable to capture effectively human perception . Akao , who is widely regarded as the father of QFD, suggested an approach involving the use of the analytic hierarchy process (AHP) . Armacost et al.  illustrated the applications of the AHP in prioritizing the CRs in an industrialized housing. Considering the vagueness and uncertainty in subjective judgments, Kwong and Bai  and  suggested the use of fuzzy AHP to determine the relative importance weights of CRs. Karsak et al.  used the analytic network process (ANP) to prioritize CRs and hence consider the interdependence among CRs. Büyüközkan et al. , Ertay et al. , and Kahraman et al.  all applied fuzzy ANP to prioritize CRs. More applications of the AHP and its variants, such as fuzzy AHP, ANP, and fuzzy ANP in assessing the relative importance weights of CRs can be found in Refs. , , ,  and . Chan et al. , Chan and Wu  and Hsu and Lin  used fuzzy and entropy methods to assess the relative importance weights of CRs. They utilized the fuzzy method to convert customers’ importance assessments of CRs into fuzzy numbers. The relative importance ratings of CRs were obtained using fuzzy arithmetic. The entropy method in the information theory was applied to analyze customers’ assessments of the performance of competitive products and to obtain the competitive priority ratings of CRs. The two sets of ratings were then combined to produce the final importance ratings of CRs. Mehdizadeh  presented a fuzzy centroid-based method for ranking CRs with competition consideration. Prasad et al.  suggested using conjoint analysis for the prioritization of CRs in the house of quality (HOQ). CRs are considered dynamic and could vary from time to time. The analysis of future CRs is critical to the long-term competitiveness of organizations, and the early prediction of future CRs can help organizations provide better products. Based on these points of view, Shen et al.  utilized fuzzy trend analysis to listen to the future voice of customers. Wu et al.  applied gray theory to analyze dynamic and future CRs. Wu and Shieh  employed a Markov chain model to analyze CRs in the future and track the importance trends of CRs from the viewpoint of probability. The assessment of the relative importance weights of CRs is a group decision-making behavior. According to the literature, the existing approaches for weight assessment all require customer preferences to be provided in the same format such as the L-point Likert scale, pairwise, or fuzzy pairwise comparison matrices. No customers can be allowed to provide a different preference format to express their preferences better. Such requirement is argued to be too demanding and may not be realistic in QFD practice. Not all customers can provide judgments in the same format because of the differences in backgrounds, education, domain knowledge, judgment capabilities, and so on. Some of them may prefer to provide their preferences in their preferred formats or those with which they are familiar. Thus, people contributing to QFD can naturally present their preferences in the formats that they prefer or are familiar with. Although the free expressions of preferences make QFD complicated, they undoubtedly make it more realistic, more flexible and more practical than before. Thus, scientific and advanced methodologies capable of handling different preference formats need to be developed to help implement QFD realistically and successfully. Our literature survey reveals that among few researchers, only Büyüközkan and Feyzioğlu  and Büyüközkan et al.  looked into multiple preference formats in QFD. Büyüközkan and Feyzioğlu  put forward a group decision-making approach based on multiple preference formats in QFD in the hope of better capturing and analyzing the CRs in software development. In their approach, customers are categorized into several focus groups, each with a different importance weight. Members within the same focus group are treated equally. Different preference formats within each focus group are first transformed into the relative importance relation matrices of the same format using different transformation functions. These relative importance relation matrices of the same format are then aggregated into a collective relative importance relation matrix using the ordered weighted geometric (OWG) operator with the fuzzy linguistic quantifier at least half. The collective relative importance relation matrix is further exploited to calculate the quantifier guided importance degrees of the CRs in a fuzzy majority sense using the OWG operator again. The quantifier-guided importance degrees of the CRs are then normalized as the relative importance weights of the CRs for the focus group. The relative importance weights of the CRs for the different focus groups are finally weighted and averaged as the final relative importance of the CRs. Büyüközkan et al.  modeled multiple preference formats in a different way and presented a different fuzzy group decision-making approach to respond better to CRs in product development with QFD. In this approach, numerical expressions are first transformed into preference relation matrices using the given transformation functions. These preference relation matrices are then converted into linguistic fuzzy preference relation matrices using a numerical-linguistic transformation function. Each linguistic fuzzy preference relation matrix is aggregated using the linguistic ordered weighted averaging (LOWA) operator with the fuzzy linguistic quantifier at least half to obtain a linguistic priority vector. The individual linguistic priority vectors are again aggregated using the LOWA operator with the fuzzy linguistic quantifier most to produce a collective linguistic priority vector for the focus group. All the collective linguistic priority vectors for the different focus groups are finally combined with the given group importance vector to produce a final linguistic priority vector, which serves as the relative importance weight vector of the CRs. Although multiple preference formats in QFD practice have been studied, the types of preference formats investigated by Büyüközkan and Feyzioğlu  and Büyüközkan et al.  are very few and insufficient. Moreover, their approaches require multiple preference formats to be transformed into the same format using pre-specified transformation functions. As the transformed information is very difficult to be precisely equalized to the original one, preference transformation may thus result in information loss or information distortion. A more natural way than information transformation is to model different preference formats without any transformation. Based on this point of view, we propose a linear goal programming (LGP) approach to assess the relative importance weights of CRs. The LGP approach does not require any information transformation and can therefore avoid information loss or information distortion. The rest of the paper is organized as follows: In Section 2, some possible preference formats that customers may use to express their preferences are introduced, and the LGP approach to modeling them together is proposed. In Section 3, a numerical example using the proposed LGP approach is examined to verify its validity, effectiveness, and potential applications in assessing the relative importance weights of CRs. Section 4 concludes the paper.
نتیجه گیری انگلیسی
Assessing the relative importance weights of CRs is a very important issue in applying QFD for new product development or quality improvement. Traditionally, the relative importance weights of CRs are assessed using an L-point Likert scale, pairwise, or fuzzy pairwise comparison matrices. No matter which method is adopted, the same preference format must be used by all the customers, and no one is exempted. Customer preferences are argued to be better captured using the formats that they prefer or with which they are familiar. Despite the fact that different preference formats make QFD more complicated, they undoubtedly make it more realistic and more practical than before. Therefore, advanced methodologies capable of handling multiple preference formats need to be developed urgently for QFD. Based on this point of view, we have proposed in this paper an LGP approach for assessing the relative importance weights of CRs in QFD. This approach integrates six possible types of preference formats, i.e., precise weights, interval weights, ordinal weights, ratio bounds, pairwise comparison matrices, and fuzzy preference relations, into one LGP model without transforming them into the same format. The LGP model is easy to implement in the Microsoft Excel Worksheet and is easy to solve using the Microsoft Excel Solver. It has been tested with an illustrative example. The testing results verify the effectiveness and practicability of the proposed LGP approach and the resultant model. Compared with the existing approaches to assessing the relative importance weights of CRs, the proposed LGP approach has three significant advantages: ♢ Six types of preference formats are discussed and modeled. Most of them have never been discussed before. There has been no attempt so far to integrate so many types of preference formats into one model. ♢ There is no need to transform different preference formats into the same format, avoiding information loss or information distortion. ♢ The proposed LGP model is linear and easy to solve.