An evidential reasoning based approach for quality function deployment under uncertainty

https://doi.org/10.1016/j.eswa.2008.06.104Get rights and content

Abstract

Quality function deployment (QFD) is a methodology for translating customer wants (WHATs) into relevant engineering design requirements (HOWs) and often involves a group of cross-functional team members from marketing, design, quality, finance and production and a group of customers. The QFD team is responsible for assessing the relationships between WHATs and HOWs and the interrelationships between HOWs, and the customers are chosen for assessing the relative importance of each customer want. Each member and customer from different backgrounds often demonstrates significantly different behavior from the others and generates different assessment results, complete and incomplete, precise and imprecise, known and unknown, leading to the QFD with great uncertainty. In this paper, we present an evidential reasoning (ER) based methodology for synthesizing various types of assessment information provided by a group of customers and multiple QFD team members. The proposed ER-based QFD methodology can be used to help the QFD team prioritize design requirements with both customer wants and customers’ preferences taken into account. It is verified and illustrated with a numerical example.

Introduction

Quality function deployment (QFD) is a cross-functional planning methodology commonly used to ensure that customer expectations or requirements, often referred to as the voice of the customer (VoC) or WHATs, are deployed through product planning, part development, process planning and production planning. It is a team-based and disciplined approach to product design, engineering and production and provides in-depth evaluation of a product. An organization that correctly implements QFD can improve engineering knowledge, productivity and quality and reduce costs, product development time and engineering changes (Besterfield, Besterfield-Michna, Besterfield, & Besterfield-Sacre, 2003). QFD has now become a standard practice by most leading organizations and has been successfully implemented world widely. A comprehensive literature review of QFD and its extensive applications is provided by Chan and Wu (2002).

The successful implementation of QFD requires a significant number of subjective judgments from both customers and QFD team members. Customers are selected for assessing the relative importance of customer expectations or requirements (WHATs). The QFD team is set up to identify customer wants, map them into relevant engineering requirements, which are often called the HOWs, meaning how the WHATs are to be met, develop the relationship matrix between WHATs and HOWs and the interrelationship matrix between HOWs, and prioritize the HOWs.

As two of the key issues of QFD, prioritization methods for WHATs and HOWs have been extensively researched and quite a number of approaches have been suggested in the QFD literature. For example, the analytic hierarchy process (AHP), a well-known and commonly used multi-criteria decision making method, and its variants:fuzzy AHP, analytic network process (ANP) and fuzzy ANP have been suggested and widely applied to prioritize customer requirements (Akao, 1990, Armacost et al., 1994, Büyüközkan et al., 2004, Ertay et al., 2005, Fung et al., 1998, Hanumaiah et al., 2006, Kahraman et al., 2006, Karsak et al., 2003, Kim et al., 2005, Kwong and Bai, 2003, Kwong and Bai, 2002, Lu et al., 1994, Park and Kim, 1998, Partovi, 2007, Wang et al., 1998). Fuzzy and entropy method (Chan et al., 1999, Chan and Wu, 2005) have also been proposed to rate the importance of customer needs. The weighted sum method (Chen and Weng, 2003, Wasserman, 1993), fuzzy weighted average (FWA) (Chen et al., 2006, Chen and Weng, 2006, Liu, 2005, Vanegas and Labib, 2001), fuzzy outranking approach (Wang, 1999) and grey model (Wu, 2006) have all been suggested for prioritizing engineering design requirements. Fuzzy logic and fuzzy inference have been extensively applied to assess the importance of WHATs and prioritize HOWs (Bottani and Rizzi, 2006, Chen et al., 2004, Chen et al., 2005, Karsak, 2004, Karsak, 2004, Khoo and Ho, 1996, Ramasamy and Selladurai, 2004, Shen et al., 2001, Temponi et al., 1999).

To reduce the heavy burden of customers and QFD team members in making their judgments, Franceschini and Rossetto (2002) develop an interactive algorithm for prioritizing product’s technical design characteristics, called IDCR (interactive design characteristic ranking), which allows the QFD team to determine a ranking order for design characteristics without using subjective rating scales and explicitly knowing the relative degree of importance of customer requirements. The IDCR algorithm avoids an inappropriate conversion from qualitative information to a relationship matrix. Han, Kim, and Choi (2004) suggest a linear partial ordering approach for assessing the information in QFD and prioritizing engineering characteristics. The linear partial information is used to extract the weights of customer wants and the relationship values between WHATs and HOWs. Four types of dominance relations that are frequently used in multi-attribute decision making with incomplete information are used to determine the priorities of engineering characteristics when the linear partial orderings of customers and QFD team members are given. The dominance relations between engineering characteristics are established through the solution of a number of linear programming models.

Considering the fact that people contributing to the QFD process may give their preferences in different formats, numerically or linguistically, depending on their backgrounds, Büyüközkan and Feyzioğlu (2005) present a fuzzy logic based group decision making approach with multiple expression formats for QFD with the hope to better capture and analyze the demand of customers, where different expressions are aggregated into one collaborative decision by using fuzzy set theory. Their approach is illustrated with a software development example. Ho, Lai, and Chang (1999) also discuss group behaviors in QFD and present an integrated group decision making approach for aggregating team members’ opinions in the case where some members in a team have an agreed criteria set while others prefer individual criteria sets. By using voting and linear programming techniques, their integrated approach consolidates individual preferences into a group consensus and is used for determining the relative importance of customer requirements.

The above literature review clearly shows that quite a lot of efforts have been made to deal with fuzziness in the process of QFD. However, no attempt has been made to address the issue of how to deal with incomplete, imprecise and missing (ignorance) information in QFD, which is essentially inherent and sometimes inevitable in human being’s subjective judgments. Fuzzy logic based approaches have been extensively used to model vagueness and ambiguity, but it cannot deal with such uncertainties as incomplete, imprecise and missing information. The purpose of this paper is to develop a rigorous and systematic methodology, on the basis of the ER approach (Wang et al., 2006b, Wang et al., 2006a, Xu et al., 2006, Yang and Singh, 1994, Yang and Sen, 1994, Yang, 2001, Yang and Xu, 2002, Yang and Xu, 2002, Yang et al., 2006), for synthesizing various types of assessment information provided by a group of customers and multiple QFD team members, which is referred to as the evidential reasoning (ER) based QFD methodology, in order to handle various types of possible uncertainties that may occur in the implementation process of QFD. The proposed ER-based QFD methodology can be used to help the QFD team to prioritize design requirements with both customer wants and customers’ preferences taken into account. It is capable of modeling various types of uncertainties using a unified belief structure in a pragmatic, rigorous, reliable, systematic, transparent and repeatable way.

The rest of the paper is organized as follows:in Section 2, we develop the ER-based QFD methodology and describe in detail its modeling mechanism and steps. The methodology is then verified and illustrated with a numerical example in Section 3. Comparisons with other QFD methodologies are provided in Section 4. The paper is concluded in Section 5.

Section snippets

The methodology

The ER approach developed for multiple attribute decision analysis (MADA) has found an increasing number of applications in recent years (Wang et al., 2006a, Wang et al., 2006b, Xu et al., 2006, Yang and Singh, 1994, Yang and Sen, 1994, Yang, 2001, Yang and Xu, 2002, Yang and Xu, 2002, Yang et al., 2006). In this section, we develop an ER-based QFD methodology to deal with various types of uncertainties in QFD. The methodology allows customers and QFD team members to express their subjective

An illustrative example

In this section, we present an illustrative example to show how the ER-based QFD methodology can be used to model uncertainty in QFD. The example is adapted from a classic QFD example about a hypothetical writing instrument (Wasserman, 1993), where easy to hold, does not smear, point lasts, and does not roll are the four identified important customer wants based on a market survey, and length of pencil, time between sharpening, lead dust generated, hexagonality, and minimal erasure residue are

Comparisons with other QFD methods

According to our literature review in Section 1, there have been no QFD methods so far that can be used for dealing with incomplete, imprecise and ignorance information in QFD. So, from the methodological point of view, the ER-based QFD methodology has significant advantages over existing QFD methods in modeling uncertainties in QFD associated with incompleteness, imprecision and ignorance.

To provide further numerical comparisons between the ER-based QFD methodology and other QFD methods, we

Conclusions

Uncertainty is inherent in human being’ subjective judgments and needs to be taken into account properly in human decision making. In this paper, we developed an ER-based QFD methodology for dealing with various types of uncertainties such as incomplete, imprecise and missing information that may occur in the implementation process of QFD. The proposed methodology allows customers and QFD team members to express their opinions using a unified belief structure independently, can accommodate

Acknowledgements

This research was supported by the Hong Kong Research Grants Council under the Grant No. CityU-1203/04E, City University of Hong Kong, SRG Project No. CityU-7002311, the UK Engineering and Physical Sciences Research Council under the Grant No. GR/S85498/01 and the UK Department of Environment, Food and Rural Affairs (DEFRA) under the Grant No. AFM222.

References (51)

  • E.E. Karsak et al.

    Product planning in quality function deployment using a combined analytic network process and goal programming approach

    Computers & Industrial Engineering

    (2003)
  • T. Park et al.

    Determination of an optimal set of design requirements using house of quality

    Journal of Operations Management

    (1998)
  • F.Y. Partovi

    An analytical model of process choice in the chemical industry

    International Journal of Production Economics

    (2007)
  • C. Temponi et al.

    House of quality: A fuzzy logic-based requirements analysis

    European Journal of Operational Research

    (1999)
  • Y.M. Wang et al.

    The evidential reasoning approach for multiple attribute decision analysis using interval belief degrees

    European Journal of Operational Research

    (2006)
  • Y.M. Wang et al.

    Interval weight generation approaches based on consistency test and interval comparison matrices

    Applied Mathematics and Computation

    (2005)
  • Y.M. Wang et al.

    Environmental impact assessment using the evidential reasoning approach

    European Journal of Operational Research

    (2006)
  • D.L. Xu et al.

    The evidential reasoning approach for multi-attribute decision analysis under interval uncertainty

    European Journal of Operational Research

    (2006)
  • J.B. Yang

    Rule and utility based evidential reasoning approach for multiple attribute decision analysis under uncertainty

    European Journal of Operational Research

    (2001)
  • J.B. Yang et al.

    The evidential reasoning approach for MADA under both probabilistic and fuzzy uncertainties

    European Journal of Operational Research

    (2006)
  • Y. Akao

    Quality function deployment: Integrating customer requirements into product design

    (1990)
  • R.L. Armacost et al.

    An AHP framework for prioritizing customer requirements in QFD: An industrialized housing application

    IIE Transactions

    (1994)
  • D.H. Besterfield et al.

    Total quality management

    (2003)
  • G. Büyüközkan et al.

    Determining the importance weights for the design requirements in the house of quality using the fuzzy analytic network approach

    International Journal of Intelligent Systems

    (2004)
  • L.K. Chan et al.

    Rating the importance of customer needs in quality function deployment by fuzzy and entropy methods

    International Journal of Production Research

    (1999)
  • Cited by (73)

    • Participatory decision-support model in the context of building structural design embedding BIM with QFD

      2018, Advanced Engineering Informatics
      Citation Excerpt :

      Fig. 3 shows a snapshot of the BIM tool with the corresponding tabs which are associated to the three HOQ steps. To conduct the numerical assessment with the stakeholders’ preferences the rating scales from Chin et al. [42] were implemented within the BIM tool to identify the corresponding PR importance weights wmn {9 = Extremely Important, 7 = Very Important, 5 = Moderately Important, 3 = Weakly Important, 1 = Very Weakly Important, 0 = Not Important}, the relationships between PR and DR Rmn {9 = Very Strong Relationship, 7 = Strong Relationship, 5 = Moderate Relationship, 3 = Weak Relationship, 1 = Very Weak Relationship, 0 = No Relationship} and the correlation of DR, rmn {9 = Very Strong Correlation, 7 = Strong Correlation, 5 = Moderate Correlation, 3 = Weak Correlation, 1 = Very Weak Correlation, 0 = No Correlation}. At Step 1, the numerical assessment (wmn) for the PR is computed, whereas the relationships (Rmn) between the PR and the DR takes place at Step 2.

    View all citing articles on Scopus
    1

    Dr. Kwai-Sang Chin is an Associate Professor of Department of Manufacturing Engineering and Engineering Management, City University of Hong Kong, China. He is the senior member of ASQ (American Society of Quality) and also the ASQ Country Representative in Hong Kong. Dr. Chin is the fellow and former Chairman of the Hong Kong Society for Quality, a world partner of ASQ.

    2

    Prof. Ying-Ming Wang is a Professor of School of Public Administration, Fuzhou University, China, and a Research Fellow of Manchester Business School, The University of Manchester, UK.

    3

    Prof. Jian-Bo Yang is a Professor of Manchester Business School, The University of Manchester, UK.

    4

    Dr. Ka Kwai Gary Poon is a Lecturer of Department of Manufacturing Engineering and Engineering Management, City University of Hong Kong, China.

    View full text