A decision model for evaluating third-party logistics providers using fuzzy analytic hierarchy process

Transcription

1 African Journal of Business Manageent Vol. 4(3), pp , March 00 Available online at ISSN Acadeic Journals Full Length Research Paper A decision odel for evaluating third-party logistics providers using fuzzy analytic hierarchy process SoonHu Soh Division of Business Adinistration, Wonkwang University, 344- Sinyong-dong, Iksan Jeonbuk, Republic of South Korea. E-ail: soonhu@wonkwang.ac.kr. Tel: Accepted Deceber, 009 As a consequence of an increasing trend toward the outsourcing of logistics activities, shippers have been faced with the inevitability of selecting an appropriate third-party logistics (3PL) provider. The selection process for the identification of a 3PL provider that best fits user requireents involves ultiple criteria and alternatives and ay be one of the ost coplex decisions facing logistics users. In this regard, this study proposes an evaluation fraework and ethodology for selecting a suitable 3PL provider and illustrates the process of evaluation and selection through a case study. It is expected that the results of this study will provide a practical reference for logistics anagers who want to engage the best 3PL provider. Future research using different datasets is warranted to verify the generalizability of the findings. Key words: third-party logistics, fuzzy analytic hierarchy process, ulti-criteria decision aking, sensitivity analysis. INTRODUCTION The third-party logistics (3PL) industry worldwide has continued its growth in recent years, and is increasing in iportance as a eans of coping with rapid changes in the global copetitive environent. As a consequence of an increasing trend toward the outsourcing of logistics activities, shippers have been faced with an inevitable need to select the best suitable 3PL provider. The use of 3PL providers can yield iportant benefits such as reduced logistics costs and fixed logistics assets, iproved order fill rates, and shortened average order-cycle lengths and cash-to-cash cycles. If an appropriate 3PL provider is not selected, serious probles can occur, such as low-quality logistics services and contract nonfulfillent. This ay then lead to the daaged reputation, iage, and trust of the shipper. Hence, the selection of a suitable 3PL provider is an iportant factor that deterines the logistics perforance. The decision aking proble for selecting the best 3PL provider has been receiving uch attention recently aong scholars as well as business practitioners. However, ost studies have siply listed selection criteria for 3PL providers through survey ethods; thus, they have not provided an integrated view. In this regard, this study proposes an integrated fraework to assist decision akers in selecting the ost appropriate 3PL provider. Evaluating and selecting 3PL providers can be regarded as a ulti-criteria decision aking (MCDM) process in which a decision aker chooses, under several selection criteria, the best option aong alternatives. One of the extensively used ethods for MCDM is the analytic hierarchy process (AHP) introduced by Saaty (980). Many researchers have applied AHP to solve MCDM probles in a nuber of different areas such as econoic planning, energy policy, project selection, and budget allocation. However, the standard AHP has often been criticized for its inability to precisely handle the inherent uncertainty or vagueness associated with the apping of a decision aker s judgent to a nuber (Chan and Kuar, 007). In any practical cases, decision akers can be iprecise about their own level of preference because of incoplete inforation or knowledge, the vagueness of the huan thought process, and the inherent coplexity and uncertainty of the decision environent. Thus, it is difficult for a decision aker to express pair wise coparison judgents as exact nuerical values on a ratio scale. To go beyond this liitation, it is ore natural to express the co-parison ratios as interval nubers or fuzzy sets because they are suitable in representing uncertain huan judgents. For this reason, this study applies a fuzzy odification of AHP (that is, fuzzy AHP) to deterine the relative iportance of selection criteria and

2 340 Afr. J. Bus. Manage. Table. Outsourced logistics services. Logistics activities 006 North Aerica Europe Asia Pacific Doestic transportation International transportation Warehousing Custos clearance and brokerage Forwarding Shipent consolidation Reverse logistics (defective, repair, return) Cross-docking Transportation anageent Product labeling, packaging, assebly, kitting Freight bill auditing and payent Supply chain consultancy by 3PL provider Order entry, processing and fulfillent Fleet anageent LLP/4PL services Custoer service Note: The data in this table were collected fro the annual third-party logistics study by C. John Langley, Jr. and research sponsors ( to eventually choose the best 3PL provider. The overall objective of this study is to establish a decision criteria fraework for evaluating 3PL providers and to propose an analytical ethod for selecting the best-suited 3PL provider. To achieve this objective, the paper is organized as follows. Firstly, in Section, soe basic concepts of 3PL are described and a review of the literature on 3PL selection criteria is provided. The fuzzy AHP ethodology is also introduced in Section 3. In Section 4, an epirical case study is presented to deonstrate the applicability of the proposed fraework and approach. Finally, Section 5 concludes with a discussion of the findings and their iplications. LITERATURE REVIEW The outsourcing of logistics activities to 3PL providers has now becoe a widely accepted practice across any industries. In fact, over 60% of Fortune 500 firs have at least one contract with a 3PL provider (Eyefortransport s USA Outsourcing Logistics Report 007). The ters of third party logistics, logistics outsourcing, and contract logistics have generally been used interchangeably both in literature and practice. 3PL refers to the utilization of external organizations to perfor all or part of the logistics activities that have traditionally been perfored within an organization (Lieb and Randall, 996). A third party is neither the shipper (first party) nor the custoer (second party) in the supply chain (Maltz and Lieb, 995). Table suarizes the specific logistics services outsourced by 3PL users. As indicated in Table, the ost frequently outsourced services in 008 were doestic and international transportation, followed by warehousing, custos clearance and brokerage, and forwarding. Most copanies have cited greater flexibility, operational efficiency, iproved custoer service levels, enhanced supply chain perforance, and better focus on their core businesses as the advantages of engaging 3PL providers (Sahay and Mohan, 006). Evaluation criteria for selecting 3PL providers To evaluate all possible supplier candidates and select the ost suitable supplier, a set of criteria ust be defined. By establishing a set of selection criteria, a copany will be better able to select a 3PL provider that will best fit its needs and existing operations (Bhatnagar et al., 999). Sink and Langley (997) presented a conceptual odel of the 3PL buying process, which is coposed of five distinct steps: () identify the need to outsource logistics, () develop feasible alternatives, (3) evaluate and select a supplier, (4) ipleent service, and (5) assess ongoing service perforance. They characterized the selection phase as an essential task in the logistics outsourcing anageent. Indeed, studies have started to provide epirical evidence supporting the iportant ipact supplier selection criteria have on the operational perforance and the overall business perforance of both the outsourcing copany and the outsourcing service provider (Kannan and Tan, 00). Evaluation criteria for selecting a 3PL provider have been

3 Soh 34 widely discussed in prior literature. The two ost frequently cited reasons for outsourcing logistics activities are cost savings and service iproveent expectations through outsourcing. Many studies have ephasized financial soundness as an essential requisite for logistics partners (Bottani and Rizzi, 006). However, cost is not the single ost iportant decision variable; logistics service issues are also considered (Selviaridis and Spring, 007). Thus, 3PL users need to balance cost with service (Setthakaset and Basnet, 005). Roberts (994) proposed the level of service provided, the quality of people, and cost as the three ost used evaluation criteria in the choice of a qualified logistics provider. This is also reflected in the work of Bhatnagar et al. (999), Dapiran et al. (996) and Lieb et al. (993); they found that cost and service represent the ost iportant criteria in logistics outsourcing decisions. Boyson et al. (999) found that financial stability, custoer service capability, and service price were rated as the ost iportant characteristic for selecting 3PL providers. In addition to cost and service, a variety of other selection criteria have been cited in prior literature. According to a survey of 54 firs offering warehousing services in the United States by Spencer et al. (994), the ost iportant evaluative criteria for selecting external or third-party service providers are, in descending order of iportance, the following: on-tie perforance, service quality, counication, reliability, service speed, and flexibility. Menon et al. (998) reported that logistics anagers consider perceived perforance and perceived capability as iportant factors in selecting 3PL providers and that these variables tend to increase in iportance when the external environent is copetitive. Perceived perforance is coprised of perceived on-tie shipents and deliveries, the ability to eet proises, the availability of top anageent, and superior error rates. Perceived capability is coprised of perceived creative anageent and the financial stability of the provider. Aghazadeh (003) provided five steps to selec-ting an effective 3PL provider and presented four relevant criteria: siilar value/objectives, up-to-date inforation technology systes, trustworthy key anageent, and a relationship of utual respect and shared willingness. In 003, the International Warehouse Logistics Associa-tion, which coprises ore than 550 logistics copanies of North Aerica, identified the following ranking of 3PL selection criteria (in a descending order): price, reliability, service quality, on-tie perforance, cost reduction, flexibility and innovation, good counication, anageent quality, location, custoize service, speed of service, order cycle tie, easy to work with, custoer support, vendor reputation, technical copetence, special exper-tise, systes capabilities, variety of available services, decreased labor probles, personal relationships, decreased asset coitent, and early notification of disruptions. Huang and Kadar (00), based on their survey of the 3PL arket in China, ranked the following criteria for the selection of 3PL providers (fro the ost to least iportant in a descending order): industry/operation experience, reputation, lower price, network coverage, own strategic asset, integrated logistics pro-viding capability, and good IT syste. Moberg and Speh (004) investigated the criteria that are considered ost iportant to U.S. warehousing custoers when selecting third party providers. According to their epirical survey, the top four selection criteria are responsiveness to service requireents, quality of anageent, track record of ethical iportance, and ability to provide value-added services. The less iportant selection criteria are (in a descending order): low cost, specific channel expertise, knowledge of arket, personal relationship with key contacts, willingness to assue risk, investent in state-ofart technologies, size of fir, and national arket coverage. The aboveentioned studies clearly show that the 3PL selection is an MCDM proble, including both quantitative and qualitative factors that are often in conflict with one another. Accordingly, this study proposes a balanced and integrated ulti-criteria hierarchical fraework for selecting 3PL providers through a careful exaination of relevant criteria. RESEARCH METHODOLOGY The selection of a 3PL provider, which is characterized by ultiple conflicting criteria, should be considered and evaluated in ters of any different criteria. AHP is one of the ost widely used ethods for addressing such MCDM probles. Although AHP is widely known as an MCDM ethod, it has often been criticized in the literature (Goodwin and Wright, 004). The ost criticized aspect is the use of a ratio scale in pair-wise coparisons, rather than an interval scale coonly used in ulti-attribute utility theory. In the conventional AHP, a pair-wise coparison is ade by using a nine-point ratio scale (Saaty, 980) to represent a decision aker s judgent or preference. Even though a crisp scale of to 9 ay be easy to apply and use, it does not explicitly take into account the aspects of uncertainty, vagueness, or fuzziness coonly inherent in huan decision aking processes. For instance, when evaluating different suppliers, due to incoplete and uncertain inforation regarding potential suppliers and their perforance, decision akers often find it is difficult to express their preferences precisely. In these situations, fuzzy set theory introduced by Zadeh (965) has been used to odel the iprecision of huan judgents. Hence, this study presents an approach that integrates the concept of fuzzy set theory with AHP to select a 3PL provider that best satisfies the deterined criteria. Many applications of the fuzzy extended AHP ethodology can be found in recent literature. Liu and Wang (009) presented an integrated fuzzy approach (fuzzy delphi, fuzzy inference, and fuzzy linear assignent) for the evaluation and selection of 3PL providers. Celik et al. (009) utilized the fuzzy AHP ethodology to odel the shipping registry selection. Cheng et al. (008) applied the fuzzy AHP ethod to calculate the relative iportance aong individual diensions and sub-criteria on the evaluation of fourth party logistics (4PLs) selection criteria. Sevkli et al. (008) integrated the AHP ethodology with the fuzzy ulti-objective linear prograing odel to solve supplier selection probles of an appliance anufacturer based in Turkey. Zhang and Feng (007) used fuzzy AHP to discuss a selection approach of reverse logistics provider

4 34 Afr. J. Bus. Manage. Table. Triangular fuzzy conversion scale. Linguistic scale Triangular fuzzy nuber Triangular fuzzy reciprocal nuber Equally iportant (,, ) (,, ) Weakly ore iportant (/3,, 3/) (/3,, 3/) Strongly ore iportant (3/,, 5/) (/5, /, /3) Very strongly ore iportant (5/, 3, 7/) (/7, /3, /5) Absolutely ore iportant (7/, 4, 9/) (/9, /4, /7).0 0 l u Figure. Triangular fuzzy nuber M. M be deonstrated as M gi, M gi,..., M gi, i =,,..., n, where all the M j gi ( j =,,..., ) are triangular fuzzy nubers. The steps of Chang s extent analysis can be given as follows: Step : The value of fuzzy synthetic extent with respect to the i-th object is defined as: S i = j= Where; M ij n M i= j= ij () through a case study. Chen et al. (006) presented a hierarchical ulti-criteria decision aking odel based on fuzzy set theory to address the supplier selection proble in supply chain anageent. So (006) proposed an approach based on fuzzy AHP and balanced scorecard (BSC) for the logistics outsourcing decision. This paper takes an approach different fro those proposed in prior literature; this study utilizes both the relative and absolute easureents for weighting a hierarchical structure and ranking the alternatives. Fuzzy analytic hierarchy process In fuzzy AHP, the triangular fuzzy nubers are used in the pair-wise coparison process to express subjective judgents. As shown in Figure, the triangular fuzzy nubers are defined by three real nubers, expressed as (l,, u). The paraeters l,, and u indicate the sallest possible value, the ost proising value, and the largest possible value that describe a fuzzy event, respectively. Their ebership functions are described as equation (). The relations between the linguistic scales and their corresponding triangular fuzzy nubers, used by experts in this study, are given in Table. 0, ( x l) ( l), µ ( x / M ) = ( u x) ( u ), 0, x < l, l x, x u, x > u. The ethodology of fuzzy AHP, based on Chang s extent analysis (99, 996), follows. Let X = { x, x,..., x n } be an object set, and G = { g, g,..., g n } be a goal set. According to the principles of Chang s extent analysis, each object is taken and extent analysis for each goal is executed, respectively. This eans that it is possible to obtain the values of extent analysis that can () M = l j= ij j j= = n M i= j= ij n n n n, j, u j, Mij = li i, ui j= j i= j= i= i= i= = n u i= i, n i= i, n l i i=.,, and Step : Since M = ( l,, u) and M = ( l,, u) are two triangular fuzzy nubers, the degree of possibility of M = l,, u ) M = ( l,, ) is defined as: V ( u ( M M ) = height( M M ) = µ ( ), = 0, ( M d l u u ) (, l ) if if if l u Where; d is the ordinate of the highest intersection point D between µ and µ (Figure ). To copare M and M, both values of M M V ( M ) M and V M M ) are required. ( Step 3: The degree possibility for a convex fuzzy nuber to be greater than k convex fuzzy M i (i=,,, k) nubers can be defined by; [( M M ) M )] V ( M M, M,..., M k ) = V and ( M M ) and... and ( M k = in V ( M M i ), i =,,..., k. (4) (3)

5 Soh 343 M M ( M M ) V l l d u u Figure. The intersection between M and M. Step 4: Assue that d( X ) inv ( S S ) i = i k for,,..., n k i. Then, the weight vector is given by k = ; T = ( d ( X ), d ( X ),..., d ( X n (5) W )) Where; X i = ( i =,,..., n) are n eleents. Step 5: Via noralization, the noralized weight vectors are: W = d( X ), d( X ),..., d( )) ( X n Where; W is a non-fuzzy nuber that gives the priority weights of one criterion over another. EMPIRICAL RESULTS In order to deonstrate a practical application of the approach presented in this paper, an epirical case study was perfored with a local anufacturing fir that had wanted to engage the services of an appropriate 3PL provider. First, a decision hierarchy for selecting a 3PL provider was designed with five levels as illustrated in Figure 3. The top level of the hierarchy stands for the ultiate goal: to evaluate and select a 3PL provider that best eets the shipper s requireents. The selection criteria and the sub-criteria, which were generated through a process of literature review and brainstoring, are shown in the second and third levels, respectively. To construct the criteria fraework, a preliinary list of criteria was prepared fro relevant literature and subsequently presented to three acadeic experts for their review to deterine the final set of candidate criteria. After an intense discussion and a round of voting, a final list of 3 criteria was deterined. During this phase, the 3 criteria were categorized into five groups: finance, service level, relationship, anageent, and infrastructure. At the fourth level, the absolute easureent ode of AHP was applied to represent the perforance of each alternative under each sub-criterion in the iediately preceding level. The use of absolute easureents T (6) rather than pairwise coparisons is recoended when a large nuber of entities are to be copared. Evaluators generally do not have enough inforation and experience to adequately consider all alternatives; as such, it is very difficult to distinguish aong and copare alternatives. Saaty (990) also pointed out that, with the absolute easureent of AHP, there can never be a reversal in the rank of the alternatives by adding or deleting other alternatives. For this study, a five-point rating scale of outstanding, good, average, fair, and poor (Liberatore, 987) was used. The lowest level contained the alternatives to be evaluated: four different 3PL providers. After constructing the decision hierarchy, pair-wise coparisons were perfored by a group of seven participants fro the case copany. The participants were requested to copare all of the ain criteria and the sub-criteria of a given ain criterion in ters of their relative iportance, using the scale given in Table. All pair-wise coparison judgents were represented as triangular fuzzy nubers in this study. As an exaple, one of the fuzzy judgent atrices of ain criteria with respect to the goal is shown in Table 3. Then, as shown in Table 4, an aggregated pairwise coparison atrix was constructed by integrating the fuzzy judgent values of different partici-pants through the fuzzy geoetric ean ethod (Buckley, 985). Buckley (985) defined the fuzzy geoetric ean r j and fuzzy weights w j of the j-th criterion fro evaluators as follows: / w r ( r r ; rj = ( a j a j), j = j ) where and represent the addition and ultiplication operations of fuzzy nubers, respectively. After obtaining the fuzzified pairwise coparison atrices, the relative iportance (weights) of all criteria and sub-criteria were calculated by the fuzzy AHP ethod described in Section.3. To identify the coputation stages clearly, the pairwise judgents fro Table 4 were evaluated as follows. Fro Table 4, the values of fuzzy synthetic extent with respect to the goal were calculated by using equation : S ( Finance ) = (3.75, 4.44, 5.36 ) ( 30.8, 5.55,.73) = ( 0.7, 0.4, 0.34)

6 344 Afr. J. Bus. Manage. L e v e l G o a l L e v e l C r i t e r i a L e v e l 3 S u b - c r i t e r i a L e v e l 4 R a t i n g s c a l e L e v e l 5 A l t e r n a t i v e s F i n a n c e L o g i s t i c s c o s t s F i n a n c i a l s t a b i l i t y R e l i a b i l i t y a n d t i e l i n e s s O u t s t a n d i n g 3 P L A S e r v i c e l e v e l Q u a l i t y o f s e r v i c e F l e x i b i l i t y a n d r e s p o n s i v e n e s s G o o d E v a l u a t i o n a n d s e l e c t i o n o f 3 P L p r o v i d e r R e l a t i o n s h i p C o p a t i b i l i t y T r u s t a n d f a i r n e s s B e n e f i t a n d r i s k s h a r i n g A v e r a g e 3 P L B 3 P L C M a n a g e e n t P e r f o r a n c e a n a g e e n t S e c u r i t y a n d s a f e t y F a i r R e p u t a t i o n a n d e x p e r i e n c e 3 P L D I n f r a s t r u c t u r e I T / I S c a p a b i l i t y L o g i s t i c s a n p o w e r P o o r Figure 3. Decision hierarchy for selecting a 3PL provider. Table 3. Triangular fuzzy judgent atrix for ain criteria. Criteria Finance Service level Relationship Manageent Infrastructure Finance (,, ) (/3,, 3/) (5/, 3, 7/) (/7, /3, /5) (5/, 3, 7/) Service level (/3,, 3/) (,, ) (/3,, 3/) (,, ) (5/, 3, 7/) Relationship (/7, /3, /5) (/3,, 3/) (,, ) (/7, /3, /5) (/5, /, /3) Manageent (5/, 3, 7/) (,, ) (5/, 3, 7/) (,, ) (3/,, 5/) Infrastructure (/7, /3, /5) (/7, /3, /5) (3/,, 5/) (/5, /, /3) (,, ) S3 ( Relationshi p) = (4.57, 5.4, 6.45) ( 30.8, 5.55,.73) = = (0.5, 0., 0.30) S 4 ( Manageent ) = (3.7, 4.36, 5.) ( = (0., 0.7, 0.4) 30.8, 5.55,.73) S5 ( Infrastruc ture) = (4.53, 5.7, 5.89) ( 30.8, 5.55,.73) = ( 0.5, 0.0, 0.7) These synthetic values were copared by using equation 3, and V ( S S) = 0. 53, V ( S S3) = 0. 7, V ( S S4) =, V ( S S5) = 0. 77, V ( S S) =, V ( S S3), V ( S S4), V ( S S5), = = = V ( S3 S) =, V ( S3 S) = 0. 8, V ( S3 S4) =, V ( S3 S5) =, V ( S4 S) = 0. 97, V ( S4 S) = 0. 50, V ( S4 S3) = 0.68, V ( S4 S5) = 0. 74, V ( S5 S) =, V ( S5 S) = 0.7, V ( S5 S3) = 0. 93, V ( S5 S4) = were obtained. Thereafter, the iniu degree of possibility was deterined fro equation 4: d ( Finance) = in (0.53, 0.7,, 0.77) = 0.53 d ( Service level) = in (,,, ) = d ( Relationship) = in (, 0.8,, ) = 0.8

7 Soh 345 Table 4. Fuzzy aggregate pairwise coparison atrix for ain criteria. Criteria Finance Service level Relationship Manageent Infrastructure Finance (.00,.00,.00) (0.57, 0.635,0.777) (0.69, 0.906,.96) (0.703, 0.944,.6) (0.88, 0.960,.9) Service level (.87,.575,.898) (.00,.00,.00) (0.703, 0.906,.78) (.57,.5,.940) (.05,.70,.343) Relationship (0.836,.04,.445) (0.849,.04,.43) (.00,.00,.00) (0.953,.04,.67) (0.934,.04,.34) Manageent (0.793,.060,.43) (0.56, 0.66,0.864) (0.789, 0.906,.049) (.00,.00,.00) (0.63, 0.73, 0.877) Infrastructure (0.886,.04,.08) (0.745, 0.855,0.985) (0.76, 0.906,.070) (.40,.369,.630) (.00,.00,.00) d ( Manageent) = in (0.97, 0.50, 0.68, 0.74) = d ( Infrastructure) = in (, 0.7, 0.93, ) = 0.7 Using equation (5), the weight vector fro Table 4 was deterined as T W = (0.53,, 0.8, 0.496, 0.7). After the noralization through equation (6), the priority weights of ain criteria were obtained as T ( 0.49, 0.8, 0.8, 0.39, 0.03). This indicates that service level is the ost iportant criteria (0.8) in selecting a 3PL provider, followed by relationship (0.8), infrastructure (0.03), finance (0.49), and anageent (0.39). By following the sae procedure, the sub-criteria were copared with respect to their correspondding ain criteria, and the priority weights of subcriteria were calculated as given in the second colun of Table 5. The global coposite priority weights given in the third colun of Table 5 were calculated by ultiplying the priority weights of sub-criteria with those of their corresponding ain criteria in the next higher level of the hierarchy. The weighting for each alternative under each sub-criterion was perfored through a rating scale technique used by Liberatore (Liberatore, 987; Liberatore et al., 99) as well as proposed in any books of Saaty (994, p.7). Table 6 shows the pair-wise coparison judgent atrix of the five-point rating scale adopted in this study. The relative weight vector was obtained by noralizing the geoetric eans for each row of the atrix, and the idealized weight vector was obtained by dividing each value of the relative weight vector by its largest value. Thus, the idealized weights of outstanding, good, average, fair, and poor were deterined as.000, 0.57,, 0.5 and 0.065, respectively. Subsequently, participants were requested to assign a rating scale to a 3PL provider with respect to each of the sub-criterion, and the resulting consensus rating scores were placed in the colun titled rating scores in Table 5. With the use of idealized weights, the ain advantage is that it prevents the undesirable rank reversal phenolenon, which potentially occurs when a new alternative is added or an existing alternative is reoved fro a set of alternatives (Rao, 007). Finally, the overall score for each of the four alternative 3PL providers was coputed for the purpose of selecting the ost appropriate 3PL provider. The overall score Si for the i-th 3PL provider was obtained using the following forula: i n j j= S = v r ij for i =,,,n, Where; v j is the global weight of j-th sub-criterion and r ij is the rating score of i-th 3PL provider with respect to j-th sub-criterion. After renoralizing the overall scores in Table 6, 3PL D was deterined to be the ost suitable alternative with respect to the shipper s requireents because it had the highest overall score (0.305) aong the four alternatives. In addition, a sensitivity analysis was perfored to test the reliability of the final result. Since the priorities and rankings of the alternatives were heavily dependent on the weights given to the decision criteria, it was necessary to exaine how sensitive the recoended decision was to the changes in the iportance of the ajor criteria (Saaty, 995). A series of sensitivity analy-ses were conducted using the Expert Choice Software. Figure 4 shows the perforance sensitivity analysis graph, where the five criteria are represented as vertical bars, the left y-axis gives the criteria iportance, and the right y-axis gives the alternative priorities with respect to each criterion. The figure shows that 3PL D outperfored all other alternatives on ost criteria, even though 3PL C perfored better on the relationship criterion and 3PL. A perfored better on the anageent criterion. The gradient sensitivity analysis in Figure 5 represents the variation of 3PL providers ranking to the changes in the relative iportance of service level as the ost influential selection criterion. 3PL D still copletely doinated all other alternatives, despite a certain degree of variation in the decision weights. Overall,

8 346 Afr. J. Bus. Manage. Table 5. Overall scores of 3PL providers. Criteria/Sub-criteria Local Weights (LW) Global Weights (GW) Rating scores 3PL A 3PL B 3PL C 3PL D GW Rating scores GW Rating scores GW Rating scores GW Finance Logistics costs Financial stability Service level Reliability and tieliness Quality of service Flexibility and responsiveness Relationship Copatibility Trust and fairness Benefit and risk sharing Manageent Perforance anageent Security and safety Reputation and experience Infrastructure IT/IS capability Logistics anpower Overall scores Renoralized scores Table 6. Pairwise coparison judgent atrix for five-point rating scale. Ratings Outstanding Good Average Fair Poor Relative vector Idealized vector Outstanding Good / Average /5 / Fair /7 /5 / Poor /9 /7 /7 / Inconsistency index = the sensitivity analysis confired the robustness of the study findings. DISCUSSION AND CONCLUSION Outsourcing has becoe a coon practice in any industries, particularly in the logistics and supply chain anageent. With ore copanies outsourcing their logistics operations, selecting appropriate and desirable 3PL providers has becoe a critical strategic decision. Nevertheless, relatively little epirical research has been undertaken on this issue, ainly because of coplexities and uncertainties inherent in the decision aking process. In this light, the present study provides a practical approach and ethodology for copanies to select a 3PL provider that best eets their requireents. In essence, the 3PL provider selection process is an MCDM proble which involves subjective value judge ents. Although AHP is the ost coon ethod for an MCDM proble, AHP sees insufficient and iprecise in ters of accurately capturing a decision aker's subject-

9 Soh 347 Figure 4. Perforance sensitivity of alternatives. Figure 5. Gradient sensitivity for service level. tive judgents regarding the interpretation of qualitative evaluation criteria. To copensate for this deficiency in the crisp pairwise coparison of the conventional AHP, a hybrid approach integrating the AHP ethodology with fuzzy logic has been proposed and applied to a practical case study for selecting the best-suited 3PL provider. The proposed fuzzy AHP approach has both the advantage of AHP, which decoposes coplex decision probles into a systeatic hierarchical structure, and the advantage of Fuzzy logic, which reflects the subjectiveness and iprecision inherent in the huan decision aking process. Thus, this study contributes by extending the fuzzy AHP

10 348 Afr. J. Bus. Manage. application to the proble of selecting a 3PL outsourcing partner, which is different fro selecting a supplier to provide a specific task or product. In addition, this study uses both the relative and absolute easureents for weighting a hierarchical structure and ranking alternatives, allowing a ore realistic and accurate representation of the 3PL provider selection proble. The ain finding of this study is the deterination of the relative iportance of selection criteria used to evaluate potential 3PL providers. Although any criteria have been proposed in prior literature for the selection of 3PL providers, less focus has been placed to their relative iportance. Hence, this study investigates the relative iportance using the fuzzy AHP approach. The results suggest inforation technology capability as the ost iportant criterion for selecting a 3PL provider, which iplies that inforation technology capability of 3PL firs is one of the ost critical factors affecting the decision of a logistics user to outsource to a 3PL provider (Lai et al., 008). This finding reflects the result of a recent survey conducted by Langley (007), which showed that inforation technology capability is one of the top three factors in the perforance of 3PL providers and one of the three ain ongoing probles with 3PL providers reported by logistics users. As such, 3PL providers should aintain a high level of inforation technology capability to develop and aintain successful logistics outsourcing relationships. This study also proposes a structured, ulti-criteria decision support odel for evaluating and selecting the best 3PL provider (Figure 3). With the developed odel, four decision choices were assued for the illustrative purpose, and the process in which an optial decision choice is ade using the fuzzy AHP approach has been explained. It is expected that this study will provide practitioners with a guide that can be used to ake better decisions when selecting 3PL providers. This study used a sall saple size, which liits the generalization of the results. However, the AHP-based approach used in this study is a subjective ethodology that perits the collection and analysis of data fro a sall group of experts (Wong and Li, 008). Nevertheless, future study is warranted to verify the results using different datasets, iproving the generalizability of the results. ACKNOWLEDGEMENTS The author would like to thank the Editor-in-Chief, Professor De la Rey van der Waldt and the anonyous reviewers for their thoughtful and helpful suggestions which helped to iprove the quality of the paper. This paper was supported by Wonkwang University in 00. REFERENCES Aghazadeh SM (003). How to Choose an Effective Third Party Logis- tics Provider. Manage. Res. News. 6(7): Bhatnagar R, Sohal AS, Millen R (999). Third party logistics services: a Singapore perspective. Int. J. Phys. Distrib. Logistics Manage. 9(9): Bottani E, Rizzi A (006). A fuzzy TOPSIS ethodology to support outsourcing of logistics services. Supply Chain Manage: Int. J. (4): Boyson S, Corsi T, Dresner M, Rabinovich E (999). Managing effective third party logistics relationships: What does it take?. J. Bus. Logistics (): Buckley JJ (985). Fuzzy hierarchical analysis. Fuzzy Sets Syst. 7(3): Celik M, Er ID, Ozok AF (009). Application of fuzzy extended AHP ethodology on shipping registry selection: The case of Turkish aritie industry. Expert Syst. Appl. 36(): Chan FTS, Kuar N (007). Global supplier developent considering risk factors using fuzzy extended AHP-based approach. OMEGA. 35(4): Chang DY (99). Extent analysis and synthetic decision. Opti. Tech. Appl. Singapore: World Sci. Iss. : Chang DY (996). Applications of the Extent Analysis Method on Fuzzy AHP. Eur. J. Oper. Res. 95(3): Chen CT, Lin CT, Huang SF (006). A fuzzy approach for supplier evaluation and Selection in supply chain anageent. Int. J. Prod. Econ. 0(): Cheng JH, Chen SS, Chuang YW (008). An Application of Fuzzy Delphi and Fuzzy AHP for Multi-criteria Evaluation Model of Fourth Party Logistics. WSEAS Trans. Syst. 7(5): Dapiran P, Lieb R, Millen R, Sohal A (996). Third party logistics services usage by large Australian firs. Int. J. Phys. Distrib. Logistics Manage. 6(0): Eyefortransport (007). USA Outsourcing Logistics Report 007. (Available fro: [Accessed: ]). Goodwin P, Wright G (004). Decision Analysis for Manageent Judgent, Third Edition. Chichester: John Wiley & Sons. Huang D, Kadar M (00). Third-Party Logistics in China: Still a Tough Market. Mercer Travel Transp. 9(): 3-0. Kannan VR, Tan KC (00). Supplier Selection and Assessent: Their Ipact on Business Perforance. J. Supply Chain Manage. 38(4): -. Lai F, Li D, Wang Q, Zhao X (008). The inforation technology capability of third-party logistics providers: a resource-based view and epirical evidence fro China. J. Supply Chain Manage. 44(3): Langley J (007). The State of Logistics Outsourcing: 007 Third-Party Logistics. Georgia Institute of Technology. (Available fro: [Accessed: ]). Liberatore MJ (987). An Extension of the Analytical Hierarchy Process for Industrial R&D Project Selection and Resource Allocation. IEEE Trans. Eng. Manage. 34(): -8. Liberatore MJ, Nydick RL, Sanchez PM (99). The Evaluation of Research Papers (Or How to Get an Acadeic Coittee to Agree on Soething). Interfaces (): Lieb RC, Millen RA, Van Wassenhove LN (993). Third Party Logistics Services: A Coparison of Experienced Aerican and European Manufacturers. Int. J. Phys. Distrib. Logistics Manage. 3(6): Lieb RC, Randall HL (996). A coparison of the use of third-party logistics services by large Aerican anufacturers, 99, 994, and 995. J. Bus. Logistics. 7(): Liu HT, Wang WK (009). An integrated fuzzy approach for provider evaluation and selection in third-party logistics. Expert Syst. Appl. 36(3): Maltz AB, Lieb RC (995). The Third Party Logistics Industry: Evolution, Drivers and Prospects. In Proceedings of the 4th Annual Transportation and Logistics Educators Conference. IL: Chicago.pp Menon MK, McGinnis MA, Ackeran KB (998). Selection Criteria for Providers of Third-Party Logistics Services: An Exploratory Study. J. Bus. Logistics. 9(): -37. Moberg CR, Speh TW (004). Third-Party Warehousing Selection: A Coparison of National and Regional Firs. A. J. Bus. 9(): 7-

11 Soh Rao RV (007). Decision Making in the Manufacturing Environent: Using Graph Theory and Fuzzy Multiple Attribute Decision Making Methods. London: Springer-Verlag. Roberts K (994). Choosing a quality contractor. Logistics Suppl. pp.4-5. Saaty TL (980). The Analytic Hierarchy Process: Planning, Priority Setting, Resource Allocation. New York: McGraw-Hill. Saaty TL (990). An exposition of the AHP in reply to the paper 'Rearks on the analytic hierarchy process'. Manage. Sci. 36(3): Saaty TL (994). Fundaentals of Decision Making and Priority Theory with the Analytic Hierarchy Process. Pittsburgh: RWS Publications. Saaty TL (995). Decision Making for Leaders: The Analytical Hierarchy Process for Decisions in a Coplex World. Pittsburgh: RWS Publications. Selviaridis K, Spring M (007). Third party logistics: a literature review and research agenda. Int. J. Logistics Manage. 8(): Setthakaset P, Basnet C (005). Third Party Logistics in Thailand - Fro the Users Perspective. In: Research Methodologies in Supply Chain Manageent, edited by Kotzab H, Seuring S, Muller M, Reiner G. Berlin: Physica-Verlag. pp Sevkli M, Koh SCL, Zai S, Deirbag M, Tatoglu E (008). Hybrid analytical hierarchy process odel for supplier selection. Ind. Manage. Data Syst. 08(): -4. Sink HL, Langley Jr. CJ (997). A anagerial fraework for the acquisition of third-party logistics services. J. Bus. Logistics. 8(): So SH (006). Fuzzy AHP Approach for the Logistics Outsourcing Decision in Supply Chain Manageent: Focused on Evaluation and Selection of Third-Party Logistics Service Providers. J. Ind. Econ. Bus. 9(4): Spencer MS, Rogers DS, Daugherty PJ (994). JIT Systes and External Logistics Suppliers. Int. J. Oper. Prod. Manage. 4(6): Wong JKW, Li H (008). Application of the analytic hierarchy process (AHP) in ulti-criteria analysis of the selection of intelligent building systes. Build. Environ. 43(): Zadeh LA (965). Fuzzy Sets. Inf. Control. 8(3): Zhang Y, Feng Y (007). A Selection Approach of Reverse Logistics Provider Based on Fuzzy AHP. In Proceeding of the Fourth International Conference on Fuzzy Systes and Knowledge Discovery - Volue 3, IEEE Coputer Society. Washington DC. pp