Determining winner in multi-attribute procurement auction: a method based on technical and business experts’evaluation information

,2,*

1.Department of Information Management and Decision Sciences,School of Business Administration,Northeastern University, Shenyang 110167,China;2.State Key Laboratory of Synthetical Automation for Process Industries,Northeastern University,Shenyang 110819,China

Determining winner in multi-attribute procurement auction: a method based on technical and business experts’evaluation information

Guangxin Gao1,Zhiping Fan1,2,and Yang Liu1,*

1.Department of Information Management and Decision Sciences,School of Business Administration,Northeastern University, Shenyang 110167,China;
2.State Key Laboratory of Synthetical Automation for Process Industries,Northeastern University,Shenyang 110819,China

A method for solving the winner determination problem (WDP)in multi-attribute procurement auctions is proposed,based on technical and business experts’evaluation information.Firstly, on the background of procurements in China,a multi-attribute procurement auction mechanism is presented,where technical and business experts participate in the bid evaluation.Then,the concept of TOPSIS is used to determine the positive and negative ideal points of the WDP according to bid prices,the technical and business experts’evaluation information.Further,the closeness coefficien of each bidder(candidate supplier)is obtained by calculating the distances to the positive and negative ideal points.Thus, the winning supplier can be determined according to the closeness coefficient.Finally,a numerical example is used to illustrate the use of the proposed method.

multi-attribute auction,winner determination,technical and business experts,bid evaluation.

1.Introduction

With the rapid development of procurement economy, reverse e-auctions or e-procurement auctions have been widely used in enterprises and governments’procurements.In the auctions,since winner determination usually depends on the bid price and other multiple attributes such as quality,service level,reputation,delivery time and so on[1–4],the auction considering multiple attributes,i.e., multi-attribute auction,is often used as a common way in procurement[2].The multi-attributeauction has manysignifican advantages,such as cost savings efficientl as well as reduction in cycle time and so on[5–8].Hence,the studies on the mechanisms or methods about the multiattribute procurement auctions have attracted many scholars’attention[3,9–13].In the real world,the attributes involved in many multi-attribute procurement auctions,in addition to the bid price,can usually be categorized into two types[13]:technical attributes(e.g.quality,after-sale service,performance)and business attributes(e.g.credit, reputation,financia situation).Usually,to evaluate the bids concerning these two types of attributes,it is necessary to invite two groups of experts,i.e.,technical experts and business experts,to participate in bid evaluation.How to develop a feasible method for determining the winner based on technical and business experts’evaluation information is a valuable research topic.

Some studies on multi-attribute auctions can be found. Thiel analyzed multi-dimensional auctions and got the conclusion that the multi-dimensional auctions are equivalent to the traditional one-dimensional auctions under the condition that the preferences of sell agency are known [14].Che developed an auction model considering quality and price to support the government procurement.In the model,each bid is evaluated by a scoring function provided by the buyer,and bidding firms costs are independent of each other[1].Following the work of Che, Branco proposed a two-stage auction model considering the case of correlation of bidding firms costs.At the firs stage a candidate fir is selected,and at the second stage the buyer bargains with the fir to reach the quality level[2].Through an experimental analysis,Bichler found that the buyer can get greater utility from the multiattributeauctioncomparedwiththesingle-attributeauction [3].Beil and Wein designed a multi-round open-ascending auction mechanism for multi-attribute auctions.In their study,an inverse-optimization-based approach was given to learn the suppliers cost functions,and a scoring rulethat maximizes the buyer’s utility was established to select the desirable supplier[15].Bichler and Kalagnanam investigated the case of multiple sourcing and developed the multi-objective integer programming models for allocation of configurabl offers[16].Teich et al.designed a multi-attribute procurement auction mechanism for multiple units of a good,where a‘pricing out’method was used to elicit buyer’s preferences about attributes and an optimization-based‘suggested price’model was proposed to calculate the best price acceptable to the buyer[9]. David et al.extended the traditional auction mechanism to the three ones,i.e.,‘first-scorsealed-bid’,‘secondscore sealed-bid’and‘sequential full information revelation’.They providedstrategies for optimal biddingprocess and maximized the buyer’s utility to determine the winner[10].Wang et al.adopted the empirical approaches to analyze the experts’bid evaluation behavior.The analysis of the cause-effect relationship shows that antagonism between technical and business expert groups tends to grow over time.Further,a new mechanism is designed to reduce antagonism between the expert groups[13].

In multi-attribute auctions,winner determination problem(WDP)is a research hotspot.Several methods have beenproposedtosolvethe WDP.Theexistingmethodscan mainly be classifie into two categories:methods based on multi-attribute decision-making(MADM)and ones based on mathematical programming models.Lai et al.developed a bid evaluation method for determining the winner,and the method is applied to the Chinese construction industry.In their method,each attribute is firs subdivided into several sub-attributes,and then hte evaluation of each attribute is obtained by integrating the scores of sub-attributes provided by experts.Further the supplier with the maximum overall score is determined as the winner[17].Kameshwaran et al.developed a multi-attribute e-procurementsystem for procuringlarge volume of a single item,and a mixed linear integer programming model was constructed to evaluate bids[18].Cheng proposed an MADMmethodto solvetheWDP inthesealed-bidreverse auction,in which the interests of both the buyer and the supplier can be satisfie[19].Singh and Benyouceftreated the WDP as an MADM problemand proposedan extended TOPSIS method for solving the problem[20].Rao et al. designeda procurementdecisionmakingmechanismof divisible goods based on multi-attribute auction,in which the winners are determined using an optimization model to maximize buyer’s utility[4].Falagario et al.adopted an extendedDEA methodto solvethesupplierselectionproblem,and the cross-efficien y evaluation was used to select the best supplier among the available candidates[21]. Yang et al.developeda linear programmingmodel to infer the buyer’s preference model.According to the preference model,winner can be determined in an intelligent framework[22].

The above methods have made significan contributions to winner determination in multi-attribute auctions from different perspectives.It should be noticed that technical and business attributes are usually considered in procurement auctions in China,and the technical and business experts are invited to provide their evaluation information with respect to technical and business attributes,respectively.However,in the existing methods,technical and business experts’evaluation information is seldom considered in multi-attribute procurement auctions.In other words,it is not suitable to solve the WDP based on technical and business experts’evaluation information using the existing methods.Therefore,it is necessary to develop a straightforward method.

This paper focuses on the WDP based on technical and business experts’evaluation information.Taking procurements of the state-ownedfirm or governmentdepartments in China as the background,a multi-attribute procurement auction mechanism is presented.In the mechanism,three kinds of attributes are considered to determine the desirable supplier,i.e.,bid prices providedby bidders,technical and business attributes.The technical and business experts are invited to provide their evaluation information with respect to technical and business attributes using scores,respectively.To determine the desirable supplier,it is necessary to aggregate the bid price and evaluation information about technical and business attributes with respect to each supplier.For this,according to the concept of classical TOPSIS,the positive and negative ideal points of the WDP are determined.Then,the closeness coefficien(the ranking value)of each candidate supplier is obtained by calculating the distances to the positive and negative ideal points.Based on the obtained closeness coefficients the winning supplier can be determined.

The remainder of this paper is organized as follows. Section 2 presents a multi-attribute procurement auction mechanism.Section 3 describes the WDP of a multiattribute auction and presents a winner determination method based on technical and business experts’evaluation information.In Section 4,a numerical example is given to illustrate the use of the proposed method.Finally, Section 5 summarizes and highlights the main features of the method proposed in this paper.

2.Multi-attribute procurement auction mechanism

In recent years,reverse e-auctions have become a common way to source a broad range of products and services [23].In the early studies,almost all the reverse auctions are conductedin a simple price format.The price-onlyfor-mat significantllimits the potential use of practical reverse auctions[22].The multi-attribute problem often occurs in business negotiations and is typical in procurement [15,24].In the procurement auction,winner determination usually depends on not only bid price but also other attributes,such as quality,service level,delivery time,payment terms and warranty terms[4].Multi-attribute auctions can promise higher market efficien y through a more effective information exchange of buyer’s preferences and supplier’s offerings[3].As extensively reported in the literature,cost saving,reduction in cycle time and guarantee of quality always accompany the use of multi-attribute auctions[5,6,8,25].Thus,as a special form of reverse auctions,the multi-attribute auction has been widely used for centralized procurement and construction project of large enterprise group and government departments[10,19].

Multi-attribute auctions are also widely used in China. With the rapid economic development in China,procurement has been undergoing a tremendous revolution in its scale and item.The procurement scale expands unceasingly and the procurement item becomes increasingly diversified In order to save cost,increase efficien y, avoid corruption and enhance the transparency of operation in procurement,many state-owned firm and governmentdepartmentsadoptthewayofmulti-attributeauctions to purchase goods.The way of a multi-attribute auction is shown in Fig.1.From Fig.1,it can be seen that one or multiple buyers,one agent,multiple suppliers and experts participate in the auction process.First,the buyers issue a request for quotations(RFQ)including basic requirements and the scoring rules for the procurement goods to the agent through the Internet.Second,the agent calls for bids to qualitative suppliers on line.Third,the suppliers formulate the bids and submit them to the agent via the Internet.After receiving the bids submitted by the suppliers,the agent organizes a referee committee composed of multiple experts to evaluate the bids.Next,the agent aggregates the experts’evaluations to determine the winner. Finally,the agentinformstheresult ofthe selectedsupplier to the buyer and suggests the buyer to sign a procurement contract with the winning supplier.

Fig.1 Way of a multi-attribute auction

In this paper,we study the WDP of the sole sourcing, i.e.,one buyer as the auctioneer and multiple suppliers as the bidders in the auction,and only one supplier can be selected as the winner finall[16].Based on Fig.1,the process of the online multi-attribute procurement auction is shown in Fig.2.The process of the auction consists of fve stages.In the following,the brief introduction of the fve stages is given.

Stage 1Invitation for bids.Usually,the buyer issues an RFQ to the agent(such as the central government procurement centre)via the Internet at the beginning of the procurement.According to the buyer’s RFQ,the agent drafts the call tender and publishes it in the website (http://www.zycg.gov.cn).Meanwhile,the agent examines thequalification ofthe supplierswhoare interestedin bidding.Then,the tender document is issued to the qualifie suppliers by email.

Stage 2Bidding for objective.On the basis of analyzing the tender document and checking their production status,thequalifie suppliers(bidders)formulatetheirbids which maximize the benefit to themselves and satisfy the buyer requirements laid out in the tender.Then,the sealed bids are submitted to the agent.Before the deadline,the submitted bids can be withdrawn,supplemented or modifie by the bidders.

Stage 3Bid opening.Once bidding for objective is closed,the agent will check all the bids submitted by the biddersandeliminatethe onesthat donotmeetthe requirements of the procurement.Then,all the candidate bidders as well as the scoring rule will be announced on the website.

Stage 4Bid evaluation.The agent organizes a referee committee composed of the multiple experts to evaluate the bids.Since technicalandbusiness attributesare considered in the process of bid evaluation,the referee committee members consist of two types of experts,i.e.,technical andbusiness experts,who are usuallyselected fromthe bid evaluation experts’database.The technical and business experts are invited to provide their evaluation information for each candidate bidder with respect to the technical and business attributes according to the scoring rule,respectively,and the experts are required to be independent of each otherduringbid evaluation.In addition,the bid prices providedby bidders in sealed bids will be transformedinto the correspondingscores by the calculation formula which is given in the scoring rule.Furthermore,the bid price’s score and the scores provided by the technical and business experts are aggregated to calculate the ranking value of each candidate bidder.In the process of calculation,the informationaboutthree kindsof attributes(technical,business and bid price)should be fully taken into account.

Fig.2 Multi-attribute procurement auction process

If the scores of the three kinds of attributes are greater for one candidate bidder,then the ranking value of the bidder will be greater,and the bidder would be determined as the winner finall.

Stage 5Award of bids.According to the obtained ranking values,a ranking of the bidders can be determined. The greater the ranking value is,the better the corresponding bidder will be.Thus,the bidder with the greatest ranking value is regarded as the winner.After the winner is determined,the result will be informed to all qualifie suppliers on the website,and the buyer will sign a purchase contract with the winner.

In this paper,we focus on Stages 4 and 5 in the process of multi-attribute procurement auction.The problem that needs to be solved is how to aggregate the bid prices and the evaluation information provided by the technical and business experts as well as how to determine the desirable supplier.

3.Method for determining winner

In this section,we give a method for solving the WDP based on technical and business experts’evaluation information.First,the notations for describing the WDP are given.Then,a winner determination method is presented.Furthermore,the procedurefordeterminingwinner is given.

3.1Notations

The followingnotations are used to represent variablesand parameters involved in the WDP in a multi-attribute procurement auction:

g:total number of bidders(bidding suppliers).

m:total number of technical attributes.

n:total number of business attributes.

p:total number of technical experts.

q:total number of business experts.

Ak:thekth bidder,k=1,2,···,g.

In the real online multi-attribute auctions,the experts are usually anonymous when the evaluation information is processed.Therefore,the weights of experts in each group (technical or business experts)should be the same.

Pk: the bid price provided by bidderAk,k= 1,2,...,g.

3.2Method

The description of the winner determination method is given below.

According to the scores concerning technical and business attributes provided by the technical and business experts,each expert’s overall evaluation value with respect to multiple attributes can be calculated,i.e.,the overall evaluation values of bidderAkprovided by the technical expertand business expert(k=1,2,...,g;s=1,2,...,p;t= 1,2,...,q),respectively.Letdenote the group evaluationvaluesofbidderAkwith respectto thetechnical attributes and business attributes,respectively,then calculation formulae ofare given by

Generally,the bid price is an important attribute for determining winner in bid evaluation.For the convenienceof analysis and calculation,the bid price needs to be transformed into the corresponding score.Letdenote the score with respect to the bid pricePk,then its calculation formula is given by

wherek=1,2,...,g.

In(3),Kis the upper limit of the score,K＞0,andEPdenotes the bench mark price,which is given by

Furthermore,a matrix can be constructed as follows:

To make the evaluation results of each bidder embody, the threeaspects(technical,business performancesand bidprice),the distances of each bidder fromV+andV?are calculated,respectively,i.e.,

wherevP,vTandvBdenote importance degrees of bid price,technical and business attributes in bid evaluation, respectively,vP,vT,vB∈[0,1],vP+vT+vB=1.Usually,they can be determined by the agent according to the buyer’s preferences on bid price,technical attributes and business attributes.For example,if the buyer regards that the bid price is important in winner determination,then the weight of the bid price is greater than the weights of the other two aspects.

whereCCk∈[0,1].CCkcan be viewed as the ranking value of bidderAk.Obviously,if bidderAkis closer to the positive ideal point and farther from the negative ideal point,namely,CCkis larger,then bidderAkis better. Thus,accordingto the obtained ranking,the desirable supplier(winner)can be selected.

Obviously,it can be seen from the above calculation process that the determination of the winner depends on the information about bid price,technical and business attributes.If the scores of one candidate bidder concerning all the three aspects(bid price,technical and business attributes)are greater,then the ranking value of the bidder will be greater,and the corresponding bidder would be determined as the winner finall.Here,Fig.3 is given to illustrate the determination of winner.Fig.3 is a threedimensional coordinate figure in which the three axes are represented by score of bid pricetechnical evaluationand business evaluationrespectively.Suppose that there are three bidders,i.e.,A1,A2andA3,and their overall evaluation results can be embodied in the form of three vectors,i.e.,OA1,OA2andOA3.The projections of the three vectors on the plane composed of the two axesrespectively.Obviously,bothare better thanIn the situation thatare about the same,we know thatA1?A2?A3since,i.e.,the bidderA1is the winner.

Fig.3 Graphical exposition of the determination of winner

In summary,the procedure for determining the winner in multi-attribute auctions is given as follows:

Step 1Let the technical and business experts provide evaluation data about each bidder with respect to technical and business attributes according to the scoring rule,k=1,2,...,g,s=1,2,...,p,t=1,2,...,q.

Step 2Calculate the group evaluation values of each bidder with regards to technical and business attributes using(1)and(2),respectively,i.e.,

Step 3Calculate the score of each bid price using(3) and(4),i.e.,

Step 4Normalizeusing(5),respectively,k=1,2,...,g.

Step 5Determine the positive ideal pointV+and negative ideal pointV?,seperately.

Step 6Calculate thedistances ofeachbidderfromV+andV?using(6)and(7),i.e.,

Step 7Calculate the relative closeness of each bidder, using(8),i.e.,CCk,k=1,2,...,g.

Step 8Determine the ranking of all the bidders according to the relative closeness and select the desirable supplier(winner).

4.Numerical example

In this section,an example on firefighti equipment procurement auction is given to illustrate the use of the proposed method.

The company Z is one of the famous state-owned petrochemical enterprises in northeastern China,which confronts with a high risk from firand explosion.To ensure the safety of staffs and state-owned property,company Z(the buyer thereafter)intends to purchase the fire fightin equipment including 100 foam extinguishers,100 dry powder extinguishers and 50 fir hydrants to substitute the old ones.According to the‘procurement management regulations of state-owned enterprises’issued by the State-owned Assets Supervision and Administration Commission,the buyer should adopt the reverse e-auction technique to procure the firefighti equipment.For this,an agent(e-procurement centre)is authorized by the buyer to organize a multi-attribute reverse auction for the procurement.At the beginningof the auction,the buyer providesa RFQ to the agent,which consists of the basic requirements of the firefighti equipments,including the detailed descriptions on technical parameters,the highest acceptable price and the delivery time.Then,the agent publishes the tender notice on the web and invites qualifie suppliers to submit sealed bids.With receiving the bids,the agent preliminarily screens the suppliers’bids according to the requirements in buyer’s RFQ.After screening,fve suppliers (A1,A2,...,A5)are determined as the candidate bidders for further bid evaluation.According to the requirements of the procurement goods,the agent considers three kinds of attributes to evaluate bids,i.e.,the bid price(Pk,Unit: China yuan),the fve technical attributesand the four business attributesThe bid prices are provided by the fve bidders,i.e.,P1= 517 236,P2=464 282,P3=406 254,P4=358 868,P5=558 211.The technical and business attributes are given in Table 1,and the scoring rule given by the agent with respect to each attribute is shown in Table 2.

Table 1 Technical and business attributes

Table 2 Scoring rule for evaluating attributes

Table 3 Scores assigned by the technical experts

Table 4 Scores assigned by the business experts

To select the most desirable supplier,the method proposed in this paper is used and the procedure is summarized below.

Firstly,using(1)and(2),the group evaluation values of each bidder with respect to technical and business attributes are calculated,and the calculation results are

Then,using(3)and(4)(K=10),the bid prices can be transformedinto the correspondingscores,i.e.,Using(5),are normalized intorespectively,k=1,2,3,4,5,andaccordingtothe winner determination matrixis constructed, i.e.,

Next,the positive and negative ideal points are determined,i.e.,andMeanwhile,according to the buyer’s preferences,the agent assigns the importance degrees of bid price,technical and business attributes respectively,i.e.,vP= 0.4,vT= 0.3 andvB=0.3.Using(6)and(7),distances of each bidderfromV+andV?can be calculated,respectively,and the calculation results are

Finally,according to the obtained closeness coefficients a ranking order of the fve bidders can be determined,i.e.,A3?A5?A4?A1?A2.Obviously,the best candidate is bidderA3.

5.Discussion and conclusions

In this paper,on the background of procurements of stateowned firm and government departments in China,a procurement auction mechanism is presented,in which thethree kinds of attributes,i.e.,bid price,technical and business attributes,are considered,and the technical and business experts are invited to participate in the bid evaluation. According to the mechanism,a method is proposed to determine the most desirable supplier.The method is based on the concept of the classical TOPSIS.The positive and negative ideal points of the bid prices,technical and business attributes are determined.The ranking value of each candidate supplier is obtained by calculating the distances to the positive and negative ideal points.According to the obtainedrankingvalues,the winningsuppliercanbe determined.Themajoradvantagesofthis paperare summarized as follows.

This paper presents a new multi-attribute procurement auction mechanism.In the mechanism,not only bid price provided by the bidders but also the evaluation informationaboutthe technicalandbusiness attributes providedby the technical and business experts is considered.The determination of the winner depends on the use of the multiattribute decision making method.This mechanism is different from the ones involved in the existing studies.It has extensive application backgrounds in China.

Thispaperpresentsa newwinnerdeterminationmethod. The method uses the information about the three kinds of attributes.It is based on the classical TOPSIS to determine the desirable supplier.The obtained result using the method simultaneously reflect the three kinds of factors: bidprice,technicalandbusiness evaluations.Theproposed methodovercomestheshortagethat thetechnicaland business attributes are not distinguished in the existing methods.

Thewinnerdeterminationmethodproposedinthispaper has the featuressuch as clear logic andsimple computation procedure.Thus,it is easy to be embedded into an online bid evaluation system.Since the method proposed in this paper is new and different from the existing methods,it gives the buyer one more choice for solving the WDP in procurement auction.

The study also has some limitations,which may serve as avenues for future research.First,the use of the decision information in the proposed method is based on the scores provided by the experts according to the scoring rules,while the scoring rules may have some impacts on the scores or the winner determination result.Thus,how to make the reasonablescoring rules accordingto the practical situation is a further noteworthy study work.Second, theproposedmethoddoes notconsiderthe subjectiveemotions,such as cooperationor antagonismbetweenthe technical and business experts,while the emotions would have impact on the winner determination result[13].Thus,how to measure the emotions and introduce the measurement results into the decision analysis of winner determination need to be investigated.Furthermore,the proposedmethod issuitabletosolvetheWDPinsingle-sourcing.Itis needed to be extended to solve the one in multiple sourcing.

[1]Y.K.Che.Design competition through multidimensional auctions.Rand Journal of Economics,1993,24(4):668–680.

[2]F.Branco.The design of multidimensional auctions.Rand Journal of Economics,1997,28(1):63–81.

[3]M.Bichler.An experimental analysis of multi-attribute auctions.Decision Support Systems,2000,29(3):249–268.

[4]C.Rao,Y.Zhao,S.Ma.Procurement decision making mechanism of divisible goods based on multi-attribute auction.Electronic Commerce Research and Applications,2012,11(4): 397–406.

[5]T.I.Tunca,S.A.Zenios.Supply auctions and relational contracts for procurement.Manufacturing&Service Operation Management,2006,8(1):43–67.

[6]W.Elmaghraby.Auctions within e-sourcing events.Production and Operations Management,2007,16(4):409–422.

[7]M.H.Rothkopf,A.B.Whinston.On e-auctions for procurement operations.Production and Operations Management, 2007,16(4):404–408.

[8]T.I.Tunca,Q.Wu.Multiplesourcing and procurement process selection with bidding events.Management Science,2009, 55(5):763–780.

[9]J.E.Teich,H.Wallenius,J.Wallenius,et al.A multiattribute e-auction mechanism for procurement: theoretical foundations.European Journal of Operational Research, 2006,175(1):90–100.

[10]E.David,R.Azoulay-Schwartz,S.Kraus.Bidding in sealedbidandEnglishmulti-attributeauctions.DecisionSupport Systems,2006,42(2):527–556.

[11]J.S.Chang,H.J.Wong.Selecting appropriate sellers in online auctions through a multi-attribute reputation calculationmethod.Electronic Commerce Research andApplications, 2011,10(2):144–154.

[12]J.Simon,F.Melese.A multiattribute sealed-bid procurement auction with multiple budgets for government vendor selection.Decision Analysis,2011,8(3):170–179.

[13]D.Wang,X.Liu,L.Liu.Bid evaluation behavior in online procurement auctions involvingtechnical andbusiness experts.Electronic Commerce Research and Applications,2013,12(5): 328–336.

[14]S.E.Thiel.Some evidence on the winner’s curse.The American Economic Review,1988,78(5):884–895.

[15]D.R.Beil,L.M.Wein.An inverse-optimization-based auction mechanism to support a multiattribute RFQ process.Management Science,2003,49(11):1529–1545.

[16]M.Bichler,J.Kalagnanam.Configurabloffers and winner determination in multi-attribute auction.European Journal of Operational Research,2005,160(2):380–394.

[17]K.K.Lai,S.L.Liu,S.Y.Wang.A method used for evaluating bids in the Chinese construction industry.International Journal of Project Management,2004,22(3):193–201.

[18]S.Kameshwaran,Y.Narahari,C.H.Rosa.Multiattribute electronic procurement using goal programming.European Journal of Operational Research,2007,179(2):518–536.

[19]C.B.Cheng.Solving a sealed-bid reverse auction problem by multiple-criterion decision-making methods.Computers and Mathematics with Applications,2008,56(12):3261–3274.

[20]R.K.Singh,L.Benyoucef.Fuzzy logic and interval arithmetic-based TOPSIS methodfor multicriteriareverse auctions.Service Science,2012,4(2):101–117.

[21]M.Falagario,F.Sciancalepore,N.Costantino,et al.Using a DEA-cross efficien y approach in public procurement tenders.European Journal of Operational Research,2012,218(2): 523–529.

[22]N.Yang,X.Liao,W.W.Huang.Decision support for preference elicitation in multi-attribute electronic procurement auctions through an agent-based intermediary.Decision Support Systems,2014,57,127–138.

[23]A.Isabelle,M.Steve,L.Annouk.Extending electronic sourcing theory:an exploratory study of electronic reverse auction outcomes.Electronic Commerce Research and Applications, 2008,7(1):119–133.

[24]B.G.Ferrin,R.E.Plank.Total cost of ownership models: an exploratory study.Journal of Supply Chain Management, 2002,38(2):18–29.

[25]T.Sandholm,D.Levine,M.Concordia,et al.Changing the game in strategic sourcing at Procter&Gamble:expressive competition enabled by optimization.Interfaces,2006,36(1): 55–68.

[26]W.Zhou,W.Jiang.Two-phase TOPSIS of uncertain multiattribute group decision-making.Journal of Systems Engineering and Electronics,2010,21(3):423–430.

Biographies

Guangxin Gaowas born in 1984.He received his B.S.degree in international business and economics and M.S.degree in business administration from Nanjing University of Science and Technology,Nanjing,China,in 2007 and 2013,respectively.He is currently working toward his Ph.D. degree at Northeastern University,Shenyang, China.His current interests include decision analysis and operations research.

E-mail:guangxin gao@126.com

Zhiping Fanwas born in 1961.He received his B.S. degree in industrial automation,M.S.degree in industrial automation,and Ph.D.degree in control theory and applications from Northeastern University (NEU),Shenyang,China,in 1982,1986,and 1996, respectively.He is currently a professor in the Department of Information Management and Decision Sciences,School of Business Administration,NEU. He was a research fellow at the City University of Hong Kong in 2001, 2003,and 2004.He is the author or coauthor of more than 70 papers published in international journals.He has been named as a Chang Jiang Scholar,China’s most prestigious scholarly honor,in recognition of his outstanding contributions to the fielof decision sciences.His current interests include decision analysis,operations research,knowledge management,and electronic commerce.

E-mail:zpfan@mail.neu.edu.cn

Yang Liuwas born in 1978.He received his B.S. degree and M.S.degree in management science and engineering from Dalian University of technology,Dalian,China,in 2001 and 2004,respectively, and his Ph.D.degree in management science and engineering from Northeastern University(NEU), Shenyang,China,in 2010.He is currently an associate professor in the Department of Information Management and Decision Sciences,School of Business Administration, NEU.He is the author or coauthor of over 10 refereed articles published in international journals includingEuropean Journal of Operational Research,Computers&Operations Research,IEEE Transactions on Systems,Men and Cybernetics,Information Sciences,andExpert Systems with Applications.His current research interests include decision analysis and operations research.

E-mail:liuy@mail.neu.edu.cn

10.1109/JSEE.2015.00137

Manuscript received October 21,2014.

*Corresponding author.

This work was supported by the National Natural Science Foundation of China(71271051;71371002;71471032)and the Fundamental Research Funds for the Central Universities,NEU,China(N140607001).