A QFD Approach for Distribution’s Location Model
Pao-Tiao Chuang
Department of Asia-Pacific Industrial and Business Management
National University of Kaohsiung
Kaohsiung, Taiwan 811, R.O.C.
Abstract
This article constructs a distribution’s location model, from the perspective of a firm’s customers, suppliers, and employees, by applying a systematic quality function deployment (QFD) approach. The proposed approach aims to assist a distribution company’s location decision in selecting an optimal location that satisfies the overall location requirements. The QFD procedure began by collecting possible candidate location requirements, followed by conducting the first stage of a sampling survey to identify the secondary location requirements. These were then sorted into major categories of location requirements. Then, the location evaluating criteria were derived from the location requirements and a central relationship matrix was established to display the degree of relationship between each pair of location requirement and location evaluating criterion. Furthermore, the second stage of a sampling survey was conducted to collect data for computing the importance weighting for each category of location requirement. During transformation of the QFD, the importance degree and the normalized importance degree of each location criterion was computed, respectively. The normalized importance degree was, finally, used as the evaluating weight in a distribution company’s location model for the analysis of location evaluation. An empirical study regarding the location decision for a distribution center in Taiwan was provided to demonstrate the proposed approach.
Keywords: Quality function deployment; Facility location decision; House of quality;
Location requirements; Distribution center.
1. Introduction
Facility location is a long-term business strategic decision for a distribution company. The strategic nature of facility location problems happens in considering either stochastic or dynamic problem characteristic [16]. Dynamic formulations focus on the difficult timing issues involved in locating a facility over an extended horizon. Stochastic formulations attempt to capture the uncertainty in problem input parameters such as forecast demand or
distance values. An appropriate facility location could perform as a good linkage role between upstream suppliers and downstream customers in the supply chain for a distribution company. The advantage of an optimal facility location is not only to reduce the transportation costs and increase the relationships between the company and its customers and suppliers, but also to improve business performance and increase competitiveness and profitability of the company.
The evaluating criteria of the facility location in most of the existing research [7][8][11][13][14][15][17][20][22] were determined from the viewpoint of the focal company. Nevertheless, a distribution’s facility location model should be constructed from the viewpoint of those who concern where the location site is and generate requirements from the facility location. That is, a facility location decision should be able to satisfy multiple objectives and needs to be determined from the viewpoints of the company’s customers, suppliers and employees.
On the other hand, quality function deployment (QFD) is a method for structured product planning and development that enables a development team to specify clearly the customer wants and needs, and then evaluates each proposed product systematically in terms of its impact on meeting those needs [4][5][10[12]. A matrix called the house-of-quality (HOQ) is used to display the relationship between the voice of the customers (WHATs) and the quality characteristics (HOWs). The HOQ is then deployed, through the QFD process, to demonstrate how the quality characteristics satisfy the customer requirements.
QFD has been proved a successful tool to support product design project that satisfies multiple customer requirements. However, the application of QFD for the location decision was rarely found in the literature. Therefore, this paper proposed applying the systematic QFD approach to determine the evaluating criteria, based on which a distribution’s location model is constructed from the viewpoint of requirements. In applying QFD to develop a distribution’s location model, the WHATs in the HOQ represent the location requirements, while the HOWs represent the location criteria. Through the QFD deployment process, the location criteria with evaluating weights were constructed for a distribution center’s location decision.
2. Literature Review
From the literature, most of the research regarding facility location problems adopted mathematical programming model or networking approach to construct an evaluation model for the facility location. With regard to a mathematical programming model, Klincewicz [15] considered construction cost, transportation cost, operating cost, and supply capacity to develop an efficient heuristic for a complex single-period facility location. Tombak [22] proposed a cost-minimization objective function for the facility location model, which caters for a local market. The model incorporates factors such as rivalry, differences between costs of domestic and offshore production, differences in consumer preferences for offshore and
domestically produced goods, and the capital costs associated with switching to domestic supply. Ghosh and Craig [11] presented a competitive equilibrium model based on the cumulative profits generated from all of the facility sites to help retailers formulate a strategic location plan in a dynamic environment. The strategic location takes into account not only the market environment confronting retailers but also anticipates possible competitive and demographic changes. The procedure involves a model for assessing site desirability, a criterion for selecting among alternative sites, and a heuristic to facilitate the computational procedure. Aikens [1] reviewed eight categories of mathematical programming models that were proposed by other researchers and have significant contributions for distribution facility location problems. The categories of mathematical model that were discussed include simple uncapacitated, simple uncapacitated multi-echelon, multicommodity uncapacitated, dynamic uncapacitated, capacitated, generalized capacitated, stochastic capacitated, and multicommodity capacitated single-echelon facility location models and problems. The solution approach for each category is discussed in terms of mathematical and computational complexity. Holmberg et al. [13] proposed a primal heuristic, which incorporates a repeated matching algorithm into the Lagrangian heuristic, for the single-source capacitated facility location problems. Holmberg [14] developed a branch-and-bound method based on a dual ascent and adjustment procedure to generate the exact solution methods for uncapacitated facility location problems where transportation costs are nonlinear and convex. In that article, an exact linearization of the costs was made, enabling the formulation of the problem as an extended, linear pure zero-one location model.
In the area of networking approach, Chen et al. [3] proposed a D.-C. programming to assist the facility location analysis for the multi-source Weber and conditional Weber problems. The D.-C. programming expresses an objective function and constraints of the facility location problem as differences of convex function. Ronnqvist [17] proposed a repeated matching algorithm to generate a feasible solution for the single-source capacitated location problem in contrast to the Lagrangian heuristics that were previously used by other researchers. The approach essentially solves a series of matching problems until certain convergence criteria are satisfied in the facility location decision. Su and Wang [20] adopted a Tabu search algorithm to develop a heuristic for location-routing problem of physical distribution. The heuristic aims to decide the optimum number of sites and the corresponding locations by considering factors such as distribution routing path and transportation costs. To assist the network analysis of facility location problem, Swink and Speier [21] presented geographic-information-system (GIS) as a decision-supporting tool for the visual display of data in the form of maps.
The evaluating criteria considered in the above work generally focus on the quantitative factors such as the construction costs of facility hardware and software, the transportation cost, and the material supply quantity. However, the selection of a facility location for a distribution company is a multi-objective decision problem. It should consider both the quantitative
(economical) and the qualitative factors. The quantitative factors include the costs of land and building, the inbound and outbound transportation costs, and the raw materials supply quantity. The qualitative factors should include the closeness to suppliers and retailers, the government policies, the environment factors, the quality of life, the availability of required technical labors and the availability of utilities. In this regard, Ross and Soland [18] provided a multicriteria approach for siting the location of public facility. Current et al. [6] reviewed the broad and multidisciplinary literature of location analysis to uncover the scope of research that has examined the multiobjective aspect of the facility location problem. Four broad categories of objective including the cost minimization, the demand-oriented objectives, the profit maximization, and the environmental concerns were discussed in the multi-objective location analysis.
In order to resolve the uncertainty problems encountered in multi-objective location analysis, Eiselt and Laporte [9] introduced the major components of a location model, in which facilities enter the market in sequential fashion. Four basic models including the main components of competitive location model, stackelberg models on a line stackelberg in a real space, stackelberg problems in network, and location games were studied. A sensitivity analysis was performed for each basic model to identify the changes of a location decision that occur when assumptions are dropped or replaced by others. Current et al. [7] developed a dynamic model for multi-objective facility location analysis that has considerable uncertainty regarding the way in which relevant parameters in the location decision will change over time. Two decision criteria, the minimization of expected opportunity loss and the minimization of maximum regret, were presented for analyzing these types of dynamic location problems, focusing on situations where the total number of facilities to be located is uncertain. Drezner and Guyse [8] utilized four rules of decision theory to examine the location problem with future uncertainty that may happen in the location evaluation factors. The four rules utilized in decision theory are the expected rule, the optimistic rule, the pessimistic rule, and the min-max regret rule. Badri [2] proposed the analytic hierarchy process (AHP) and the multi-objective goal-programming methodology as aids in making location-allocation decisions to solve the volatile and complex global facility location problem.
3. A Basic Structure of HOQ for Distribution’s Location
The basic structure of a HOQ for developing a distribution company’s location model is depicted in Figure 1. As shown in the figure, a typical HOQ contains the following five sections.
1. A structured list of the location requirements, which are the quality requirements in the traditional HOQ for product design, to represent the needs of the distribution company’s customers, suppliers, and employees to the facility location.
importance degree of each location requirement.
3. The location criteria, which are the quality characteristics in the traditional HOQ for product design, to represent the evaluating criteria that should be considered to satisfy the location requirements.
4. A central relationship matrix to link the relationship between the location requirements and the location criteria. The central relationship matrix displays the degree to which each location criterion satisfies the corresponding location requirement.
5. A row vector named the importance degree of location criteria to identify the degree to which each location criterion satisfies the overall location requirements and a row vector named the normalized importance degree of location criteria to represent the relative importance and the resource allocation priority among these location criteria. The normalized importance degree of each criterion is finally used as the evaluating weight in the distribution company’s facility location model.
(Take in Figure 1)
4. A QFD Process for Distribution’s Location
The planning procedure for a QFD process to develop a distribution company’s location model is expressed step by step as follows.
1. The planning procedure begins by identifying what location requirements need to be satisfied. In this regard, this paper suggests the location planning team investigating the voice of the distribution company’s customers, suppliers, and employees about their requirements to the facility location. The collected information is formed as the candidate requirements listed in a questionnaire for sampling survey of those who have needs to the distribution company’s location about their opinion whether each requirement should exist. After a statistical significance test, the confirmed requirement items are identified as the secondary location requirements, which are further sorted into major categories of location requirements in the HOQ for the QFD process by using an affinity diagram (KJ method).
2. In the second step, the location criteria are developed by the location planning team. The location criteria are derived from the secondary location requirements to express what factors should be considered in order for the distribution’s location that satisfies the location requirements. The location criterion is a term used as the internal and technical language of a distribution company and is placed on top of the HOQ.
3. A central relationship matrix is established to display the degree of relationship between each location requirement and the corresponding location criterion. This matrix should be constructed by integrating cross-functional expert knowledge of the distribution’s location
planning team. In the central relationship matrix, a symbol representing strong, moderate, or weak relationship in each cell reflects the extent to which a particular location criterion contributes to meet the corresponding location requirement. During the transformation, quantified scales, such as 4-2-1 or 9-3-1, are used to denote strong-moderate-weak relationship for the QFD computation.
4. The column vector of the importance weighting of location requirement is the place to record the degree of importance that each location requirement is perceived by those who concern where the distribution location site is. The distribution’s location planning team can conduct a sampling survey or investigation, in which the company’s customers, suppliers, and employees are asked to rate the importance degree of each requirement based on a scaled system provided in the questionnaire. The number of points on such a scaled system has been known to range from three to ten. This paper suggests using a Likert five-scale, where the values 1,2, 3, 4, 5 are used to denote “Very unimportant”, “Unimportant”, “Neutral”, “Important”, and “Very important”, respectively, to survey the opinion of the respondents. The sampling data is then used to compute the average importance degree for each location requirement, which constructs the column vector of the importance weighting of requirement.
5. Finally, the importance degree of each location criterion is computed from the weighted column sum of the importance weighting of each requirements multiplied by the quantified relationship value of the corresponding location criterion in the central relationship matrix. That is, if n location criteria are considered for the purpose of satisfying m location requirements, the importance degree of each location criterion is computed as Equation (1).
∑
= ⋅ = m i i ij j R c w 1 (1)Where, wj = The importance degree of the jth location criterion, j = 1, 2,….., n.
ij
R = The quantified relationship value between the ith location requirement and the jth location criterion in the central relationship matrix.
i
c = The importance weighting of the ith requirement, i = 1, 2, …, m.
The importance degree of each criterion is then normalized to a total of 1 to represent the evaluating weight in the distribution’s facility location model. The normalized process [23] for each location factor is shown below.
∑ = = n j j w j w j w Normalized 1 (2)
5. Empirical Study
To demonstrate how the proposed approach constructs the distribution company’s location model, this research performs an empirical study regarding to the location decision of a distribution center (DC) in Taiwan. In the example, three candidate locations are considered for further evaluation.
5.1 Development of location requirements and location criteria
To develop location requirements and criteria for the distribution center’s location decision, this research, first, reviewed the location criteria used in the existing research from which lists of possible location requirements were compiled. Then, the QFD process started from interviewing the voices of the DC’s customers (retailers), suppliers, and employees to collect possible candidate requirements. In the interview process, the compiled lists of possible location requirements were provided to those who participated the interview. To ensure collectively exhaustive, the participants were allowed to add, delete, or modify any item on the lists of location requirements according to their practical experiences. Through this process, 34 candidate requirements are gathered in a questionnaire for the first stage of sampling survey. In this stage, the purpose of the sampling survey is to confirm which items should be on the lists of secondary location requirements. In the survey, each respondent answers「Yes」or「No」to reflect his/her opinion about whether a particular candidate requirement should exist as a secondary location requirement. The sampling survey mailed 200 questionnaires to the respondents. Among them, 70 were the DC’s customers, 70 were the DC’s suppliers, and 60 were the DC’s employees. Finally, 76 questionnaires were received and believed to be valid for further analysis. Collected sample data was run by a statistical test for each particular candidate requirement to confirm if it is necessary to be considered in the secondary location requirements. The hypothesis test is shown as Equation (3).
> ≤ 5 . 0 : 5 . 0 : 0 i a i p H p H (3) Where, pirepresents the proportion of the opinion that the ith candidate requirement should be listed
on the secondary location requirement.
The statistical testing results are shown in Table I. From Table I, 23 candidate requirements show statistically significant. Therefore, these 23 candidate requirements were confirmed as the secondary location requirements, which were then sorted into ten major categories of location requirement by adopting an affinity diagram (see Table II). In addition, also shown in Table II, the secondary quality characteristics were then derived from the location requirements. And the derived secondary quality characteristics were further sorted into nine major dimensions of quality characteristics, which are shown in Table III. These quality characteristic dimensions represent the location criteria that should be assessed in
order to satisfy the location requirements.
(Take in Table I)
(Take in Table II)
(Take in Table III)
5.2 Establishment of central relationship matrix
To establish the central relationship matrix, this research integrates cross-functional expert knowledge of the DC's location planning team to identify the relationship degree between each pair of location criterion and the corresponding location requirement. Each relationship degree such as strong, moderate, weak, or no relationship in the matrix represents how a particular location criterion satisfies the corresponding location requirement. An ordinal symbol「」,「」,「」or blank is adopted in the matrix to represent「strong」, 「moderate」,「weak」or no relationship, respectively. During transformation of the QFD, quantified scale 9-3-1-0 was used to denote strong-moderate-weak-no relationship for computation purpose. As shown in Table IV, for instance, to satisfy the location requirement 「fast/precise distribution ability」, the location criterion 「land features」 contributes as a weak relationship, thus 「」 is shown in the corresponding cell of the matrix. Moreover, the 「initial and operating costs」 contributes as a strong relationship「」 and the 「closeness to suppliers and retailers」 contributes as a moderate relationship「」; whereas, the 「community and working environment」 has no relationship with「fast/precise distribution ability」, therefore, the corresponding cell is blank.
(Take in Table IV)
5.3 Importance weighting of location requirement
To determine the importance weighting of location requirement, the second questionnaire, involving ten location requirement categories sorted in Table II, was designed to survey the importance degree of each category. The same respondents in section 5.1 were asked to identify the importance degree, valued from 1 to 5, of each location requirement category. The higher the value is the more important the particular category, the lower the less important. In this stage of the sampling survey, a follow-up telephone-call was made to each respondent; therefore 92 out of 200 mailed questionnaires were received and believed to be effective sample data. The collected data was used to compute the average value, shown in Equation (4), of the importance degree for each location requirement category. The computed results were shown on the last column of Table IV to represent the importance weighting of
each location requirement category for QFD deployment process. k x c k j ij i
∑
= = 1 (4) Where, ic = The importance weighting of the ith location requirement category, i = 1, 2, …, 10. ij
x = The jth sample data value of the importance degree for the ith location requirement category. k = Number of effective sample data. In this example, k= 92.
5.4 Importance degrees of location criteria
According to Equations (1) and (2), the importance degree and the normalized importance degree of each location criterion were computed, respectively. The normalized importance degree for each location criterion was then used as the evaluating weight in the DC’s facility location model. The computed results are shown on the bottom two rows of Table IV. For instance, the importance degree of the 「 land features 」 =1×4.51+9×4.00+9×4.34+9×3.90+9×4.37=154.0 ; and the normalized importance degree of the
「land features」= 0.116 2 . 110 6 . 31 3 . 99 9 . 24 0 . 145 5 . 139 8 . 215 8 . 297 0 . 154 0 . 154 = + + + + + + + + .
Table IV also shows that the resource allocation priorities of the location criteria follow the sequence: initial and operating costs >> transportation conditions >> land features >> political regulation and law >> closeness to suppliers and retailers >> information technology conditions >> labor conditions >> energy/utilities >> community and working environment (“>>” means “more important than”). This could provide the company the advice of the resource allocation policy to improve the conditions of those criteria that are more important, when the DC tries to increase its competitive advantage of the location.
5.5 Location decision for the example
From the computed normalized importance degree of each location criteria in previous subsection, the DC’s location model with the evaluating weight for each location criteria was constructed as Table V. In this case, three candidate locations, with one in northern Taiwan and two in southern Taiwan, for the DC were selected for further evaluation based on this location model. It is noted that the evaluating score for a particular location criterion in each candidate location is evaluated based on the location site’s condition-level, from worst to best, in that criterion. The highest evaluating score is 100 and the lowest is 0 in each particular location criterion. Finally, the cumulated weighting scores were computed by Equation (5) for the purpose of determining which one was the optimal selection. The highest cumulated
weighting score for each candidate location is 100. The higher the score is for a given candidate location the more the possibility that the location decision will select it. From Table V, the candidate location B, located in southern Taiwan, has the highest cumulated weighting score 82.8. Therefore, the facility location decision would select the candidate location B for the DC’s new location site.
∑
= = n i ij i j we S 1 (5) Where, jS =The cumulated weighting score for the jth candidate location site. j=1,2,……
i
w =The evaluating weight of the ith location criterion, i=1,2,…,n.
=The normalized importance degree of the ith location criterion from Equation (2).
ij
e =The evaluating score of the jth candidate location on the ith location criterion. n =Number of location criteria that are considered. In this location model, n=9.
(Take in Table V)
6. Discussion of Findings
This paper applied the QFD to construct a facility location model, based on which a DC company in Taiwan selected its optimal location site from three candidate locations. The QFD based facility location model provides the company an opportunity to make the location decision from the perspective of location requirements.
In constructing a location model, this research used two stages of sampling survey to confirm the candidate requirements that should be considered and to identify the importance weighting of each location requirement category, respectively, in the QFD process. To ensure the candidate location requirements were collectively exhaustive in the first stage of sampling survey, the participants were allowed to add, delete, or modify any item on the preview lists of location requirements according to their practical experiences. In the second stage of sampling survey, the respondents were asked to identify the importance degree, valued from 1 to 5, of each location requirement category. The sampling data showed that the respondents tended to perceive almost every requirement as being important, rating from 4 to 5. Fortunately, this did not harm to the purpose of this research because the location planning team could still differentiate the relative importance of location requirement from each other.
On the other hand, to establish the central relationship matrix, this research integrates cross-functional expert knowledge of the DC's location planning team to identify the
relationship degree between each pair of location criterion and the corresponding location requirement. This implies that the company needs to form a cross-functional location planning team in order to make an objective and optimal location decision that satisfies the overall location requirements.
In addition, during transformation of the QFD, this paper used the quantified scale 9-3-1-0 to denote strong-moderate-weak-no relationship for computation purpose. The facility location practitioners could also use other quantified scale, such as 4-2-1-0, to denote the relative degree of relationship. In this case, the ratio of the quantified value of strong relationship to that of moderate relationship and the ratio of the quantified value of moderate relationship to that of week relationship should be identical. And the ratio is proportional to that of 9-3-1-0. That is, if 4-2-1-0 quantified scale is used, the ratio of relative degree of relationship is 4/2=2/1=2. It is proportional to the ratio of 9-3-1-0, which is 9/3=3/1=3. The use of different scale system would only affect the value of the importance degree of each location criteria. The normalized importance degree of each location criteria would be the same as those shown on the bottom row of Table IV.
Finally, in section 5.4, this research found that the 「initial and operating costs」and the 「transportation conditions」location criteria were more important than others. This implied that the costs and the transportation issues still played very important roles in the distribution’s location decision. Also from section 5.4, the 「 information technology conditions」location criterion was medium important. This is because the empirical study was performed by the end of year 1999. Whereas, the information technology to the distribution center has been getting more and more important in the trend of global supply chain environment. Therefore, a larger evaluating weight for the 「 information technology conditions」 location criterion might be expected if the proposed approach is applied in today’s supply chain facility location decision.
7. Conclusions
The selection of a distribution company’s location is a multi-objective decision problem. It should consider both the quantitative and the qualitative factors. In addition, an optimal location should be selected to satisfy those who concern where the distribution’s location site is. That is, a distribution’s location model should be constructed from the viewpoints of those who have needs to the company’s location. Thus, the location decision of a distribution company should be requirements-driven.
In this regard, the QFD is a method for structured product planning and development that enables a development team to specify clearly the customer wants and needs, and then evaluates each proposed product systematically in terms of its impact on meeting those needs. Though the QFD technique has been proved to be a successful tool to support a product design project, the application of QFD to the location decision was rarely found in the
literature. Therefore, this paper proposed a QFD approach to construct a distribution’s location model from the viewpoints of the company’s customers, suppliers, and employees. An empirical study regarding the location decision for a distribution center in Taiwan was provided to demonstrate the proposed approach. This approach could provide a distribution company an objective perspective in making the location decision and satisfying the location requirements of the company’s customers, suppliers, and employees.
Due to the vagueness of the colloquial and semantic information of the location requirements and the degree of relationship in the central relationship matrix, future research may apply the fuzzy theory to assist the proposed QFD approach in solving the problem faced in this situation. In addition, to better differentiate the importance of each location requirement, the analytic hierarchy process (AHP) developed by Saaty [19] could be combined to the proposed QFD approach in a further research. Finally, to solve the resource-constraint problems that are encountered in today’s supply chain environment, future research can incorporate the proposed approach with the capacitated multi-echelon problems to make a better facility location decision.
References
1. Aikens, C.H. (1985), “Facility Location Models for Distribution Planning”, European Journal of Operational Research, Vol.22, pp.263-279.
2. Badri, M.A. (1999), ”Combining the Analytic Hierarchy Process and Goal Programming for Global Facility Location---Allocation Problem,” International Journal of Production Economics, Vol.62, No.3, pp.237-248.
3. Chen, P.C., Hansen, P., Jaumard, B., and Tuy, H. (1998), “Solution of the Multisource Weber and Conditional Weber Problems by D.-C. Programming,” Operations Research, Vol.46, No.4, pp.548-562.
4. Chuang, P.T. (1999), “The Application of Fuzzy Reasoning for Quality Function Deployment,” Journal of the Chinese Institute of Industrial Engineers, Vol.16, No.6, pp.725-732.
5. Cohen, L. (1995), Quality Function Deployment
How to Make QFD Work for You, Addison-Weiley Publishing Co., New York.6. Current, J., Min, H., and Schilling, D. (1990), “Multiobjective Analysis of Facility Location Decisions,” European Journal of Operational Research, Vol.49, pp.295-307. 7. Current, J., Ratick, S., and Revelle, C. (1998), “Dynamic Facility Location When the Total
Number of Facilities Is Uncertain: A Decision Analysis Approach,” European Journal of Operational Research, Vol.110, No.3, pp.597-609.
8. Drezner, Z. and Guyse, J. (1999), “Application of Decision Analysis techniques to the Weber Facility Location Problem,” European Journal of Operational Research, Vol.116, No.1, pp.69-79.
9. Eiselt, H.A. and Laporte, G. (1996), “Sequential Location Problems,” European Journal of Operational Research, Vol.96, pp.217-231.
10.Franceschini, F. and Rossetto, S. (1995), “QFD: The Problem of Comparing Technical/Engineering Design Requirements.” Research Engineering Design, Vol.7, pp.270-278.
Environment”, Journal of Marketing, Vol.47, pp.56-68.
12.Hauser, J. R. and Clausing, D. (1988), “The House of Quality”, Harvard Business Review, May-June, pp.63-73.
13.Holmberg, K., Ronnqvist, M., and Yuan, D. (1999), “An Exact Algorithm for the Capacitated Facility Location Problems With Single Sourcing,” European Journal of Operational Research, Vol.113, No.3, pp.544-559.
14.Holmberg, K. (1999), “Exact Solution Methods for Uncapacitated Location Problems With Convex Transportation Costs,” European Journal of Operational Research, Vol.114, No.1, pp.127-140.
15.Klincewicz, J. G. (1985), “A Large-Scale Distribution and Location Model”, AT&T Technical Journal, Vol.64, pp.1705-1730.
16.Owen, S.H. and Daskin, M.S. (1998), “Strategic Facility Location: A Review,” European Journal of Operational Research, Vol.111, No.3, pp.423-447.
17.Ronnqvist, M. (1999), “A Repeated Matching Heuristic for the Single-Source Capacitated Facility Location Problem,” European Journal of Operational Research, Vol.116, No.1, pp.51-68.
18.Ross, G. T. and Soland, R. M. (1980), “A Multicriteria Approach to the Location of Public Facilities,” European Journal of Operational Research, Vol.4, pp307-321.
19.Satty, T. L. (1988), “The Analytic Hierarchy Process,” RWS Publications, Pittsburgh. 20.Su, C.T. and Wang, C.Y. (1998), “A Tabu Search Heuristic for the Location-Routing
Problem,” Journal of Commercial Modernization (Taiwan), Vol.1, No.1, pp.107-123. 21.Swink, M. and Speier, C. (1999), “Presenting Geographic Information: Effects of Data
Aggregation, Dispersion, and Users’ Spatial Orientation,” Decision Science, Vol.30, No.1, pp.169-195.
22.Tombak, M. M. (1995), “Multinational Plant Location As a Game of Timing,” European Journal of Operational Research, Vol. 86, pp.434-451.
23.Wasserman, G. S. (1993), “On How to Prioritize Design Requirements During the QFD Planning Process.” IIE Transactions, Vol.25, No.3, pp.59－65.
Figure 1. An HOQ for location model. Location criteria (Quality characteristics) Central relationship matrix L o ca ti o n r eq u ir em en ts (Q u al it y r eq u ir em en ts ) Im p o rt an ce w ei g h ti n g o f re q u ir em en t
Importance degree of location criteria
Table I. Significance tests for location requirements. Candidate Requirement p Z Signifi -cance Candidate Requirement p Z Signifi -cance
1.No traffic jam during delivery 0.605 1.878 ** 18.Packing in accordance with
customer needs 0.513 0.230
2.Discount price for the goods 0.526 0.459 19.Safe public utilities 0.816 7.102 ** 3.Short delivery time 0.750 5.033 ** 20.Quick response to customer
needs 0.645 2.636 **
4.Available goods for sale 0.355 -2.636 21.Reduce inventory level for goods 0.355 -2.636 5.Good service 0.566 1.157 22.Transportation infrastructure
impact 0.658 2.902 **
6.Fast information supply 0.684 3.455 ** 23.Siting on flat land 0.697 3.746 ** 7.Near main reads 0.658 2.902 ** 24.Easily expandable 0.737 4.688 ** 8.Close to high-population area 0.618 2.125 ** 25.Labor availability 0.908 12.297 ** 9.Enough parking lots 0.868 9.500 ** 26.Low employee turnover 0.816 7.102 ** 10.Close to suppliers 0.697 3.746 ** 27.No shortage on electricity
and water supply 1.000 99999 ** 11.Lower goods price 0.553 0.922 28.Close to retailers 0.934 15.267 ** 12.Fair transportation fare 0.961 20.611 ** 29.Convenient on- and off- duty for
employees 0.855 8.805 ** 13.Frequent delivery 0.882 10.297 ** 30.Standard specification of goods
and pallet 0.500 0.000 14.Good quality for delivery goods 0.579 1.393 31.Number of work/vacation days 0.316 -3.455 15.Close to financial institute 0.237 -5.397 32.Easily inbound transportation 0.895 11.210 ** 16.Close to government agency 0.171 -7.618 33.Easily outbound transportation 0.816 7.102 ** 17.Short delivery lead time 0.711 4.046 ** 34.More branch sites 0.658 2.902 **
Table II. Development of location requirement categories and derivation of quality characteristics.
Derivation of quality characteristics Secondary location
requirements Location requirement Derived to Secondary quality categories characteristics
Short delivery time Frequent delivery Short delivery lead time
Fast/precise distribution ability
Sector of an area, transportation cost, supply speed, product feature, downstream retailers’ locations, employees’ professional ability,
telecommunication infrastructure.
No traffic jam during delivery Near main roads Transportation infrastructure
Impact
Convenient on- and off- duty for employees
Convenient transportation
Facility site, land price, construction cost, vehicle flow, road width, future transportation
development, geographical feature, telecommunication infrastructure, construction
regulation, industry policy.
Enough parking lots Siting on flat land Easily expandable
Adequate land space
Parking space, land price, construction cost, geographical feature, surrounding environment, construction regulation, industry
policy, environment regulation.
Close to suppliers Close to retailers More branch sites
Appropriate site location
Sector of an area, facility site, land price, transportation distance, future transportation development, upstream suppliers’ locations,
downstream retailers’ locations.
Fair transportation fare Low transportation cost
Sector of an area, transportation cost, vehicle flow, transportation distance, transportation mode and system, product feature, upstream suppliers, locations, downstream retailers’ locations, telecommunication infrastructure.
Easily inbound transportation Easily outbound transportation
Smooth inbound/outbound transportation
Parking space, sector of an area, facility site, construction cost, transportation mode and
system, road width, future transportation development, geographical feature, product
feature, construction regulation, industry policy.
Close to high-population area Labor availability Low employee turnover
Available labor supply
Salary/wage, transportation mode and system, future transportation development, geographical feature, product feature, work
safety, work environment, surrounding environment, part-time labor availability, welfare/benefits, employees’ professional
ability, industry policy.
Safe public utilities Available public facilities
Sector of an area, construction cost, geographical feature, construction regulation,
industry policy, work safety, work environment, surrounding environment.
No shortage on electricity
and water supply Stable utilities supply
Sector of an area, construction cost, stable energy supply, industry policy, environment
regulation.
Fast information supply Quick response to
customer needs
Quick response to requirements
Telecommunication infrastructure, information formats standardization, degree of software
Table III. Development of location criteria. Secondary quality characteristics Quality characteristic dimensions (Location criteria) Secondary quality characteristics Quality characteristic dimensions (Location criteria)
Parking space Construction regulation
Sector of an area Industry policy
Facility site
Land features
Environment regulation
Political regulation and law
Land price Work safety
Transportation cost Work environment
Salary/wage Surrounding environment Community and working environment Construction cost
Initial and operating costs
Part-time labor availability
Vehicle flow Employees’ professional
ability
Transportation distance Welfare/benefits
Labor conditions
Transportation mode
and system Stable energy supply Energy/utilities
Supply speed Telecommunication
infrastructure
Road width Information format
Standardization Future transportation Development Transportation conditions Degree of software application Information technology conditions Geographical feature Product feature Upstream suppliers’ locations Downstream retailers’ locations Closeness to suppliers and retailers
Table IV. The QFD process for the empirical study. Location criteria L an d f ea tu re s In iti al a n d o p er at in g co st s T ra n sp o rta tio n co n d iti o n s Cl o se n es s t o s u p p lie rs an d r et ai le rs P o lit ic al r eg u la tio n a n d L aw Co m m u n ity a n d w o rk in g en v ir o n m en t L ab o r c o n d iti o n s E n er g y / u til iti es In fo rm at io n te ch n o lo g y co n d iti o n s Im p o rt a n ce w ei g h tin g o f r eq u ir em en t
1.Fast/precise distribution ability 4.51
2.Convenient transportation 4.00
3.Adequate land space 4.34
4.Appropriate site location 3.90
5.Low transportation cost 4.37
6.Smooth inbound/outbound
Transportation 4.37
7.Available labor supply 3.95
8.Available public facilities 3.73
9.Stable utilities supply 3.51
L o ca tio n r eq u ir em en t c a te g o ri es 10.Quick response to requirements 4.44
Importance degree of location
criteria 154.0 297.8 215.8 139.5 145.0 83.8 99.3 85.1 105.7 Normalized Importance degree 0.116 0.225 0.163 0.105 0.109 0.063 0.075 0.064 0.080 ( Quantified value of relationship: strong： = 9 ； moderate： = 3 ； weak： = 1 ; no relationship: blank=0)
Table V. The location decision model for the example
Evaluating score for candidate location
Location criteria Evaluating weight Location site condition-level 0 ………100 Candidate location A Candidate location B Candidate location C
Land features 0.116 worst ….…..best 70 77 100
Initial and operating costs 0.225 worst .……. best 100 100 70
Transportation conditions 0.163 worst ….… best 72 55 55
Closeness to suppliers and retailers 0.105 worst ….….best 30 65 30
Political regulation and law 0.109 worst ……. best 50 100 30
Community and working environment 0.063 worst ……. best 50 100 100
Labor conditions 0.075 worst ……. best 75 75 75
Energy / utilities 0.064 worst ……. best 70 100 100
Information technology conditions 0.080 worst ……. best 60 80 50
