INFORMATICA,2019,Vol.30,No.1,135–152 135 2019 VilniusUniversity DOI:http://dx.doi.org/10.15388/Informatica.2019.201
ABipolarFuzzyExtensionoftheMULTIMOORA Method
DragisaSTANUJKIC1 ∗,DarjanKARABASEVIC2 , EdmundasKazimierasZAVADSKAS3,FlorentinSMARANDACHE4 , WillemK.M.BRAUERS5
1TechnicalFacultyinBor,UniversityofBelgrade,Serbia
2FacultyofAppliedManagement,EconomicsandFinance, UniversityBusinessAcademyinNoviSad,Belgrade,Serbia
3InstituteofSustainableConstruction,LaborofOperationalResearch, FacultyofCivilEngineering,VilniusGediminasTechnical University, Sauletekio11,Vilnius,LT-210233,Lithuania
4DepartmentofMathematics,UniversityofNewMexico, 705GurleyAvenue,Gallup,NM87301,USA
5FacultyofBusinessandEconomics,DepartmentofEconomics, UniversityofAntwerp,Antwerp,Belgium
e-mail:dstanujkic@tfbor.bg.ac.rs,darjan.karabasevic@mef.edu.rs,edmundas.zavadskas@vgtu.lt, fsmarandache@gmail.com,willem.brauers@uantwerpen.be
Received:October2018;accepted:February2019
Abstract. TheaimofthispaperistomakeaproposalforanewextensionoftheMULTIMOORA methodextendedtodealwithbipolarfuzzysets.Bipolarfuzzysetsareproposedasanextension ofclassicalfuzzysetsinordertoenablesolvingaparticularclassofdecision-makingproblems. Unlikeotherextensionsofthefuzzysetoftheory,bipolarfuzzysetsintroduceapositivemembership function,whichdenotesthesatisfactiondegreeoftheelementxtothepropertycorrespondingto thebipolar-valuedfuzzyset,andthenegativemembershipfunction,whichdenotesthedegreeofthe satisfactionoftheelementxtosomeimplicitcounter-propertycorrespondingtothebipolar-valued fuzzyset.Byusingsingle-valuedbipolarfuzzynumbers,theMULTIMOORAmethodcanbemore efficientforsolvingsomespecificproblemswhosesolvingrequiresassessmentandprediction.The suitabilityoftheproposedapproachispresentedthroughanexample.
Keywords: bipolarfuzzyset,single-valuedbipolarfuzzynumber,MULTIMOORA,MCDM.
1.Introduction
Themanagementofverycomplexsystemsisthemostcomplex,andthereforethemostdifficulttaskofthemanagersoftoday’sorganizations.Theeffectivenessofthemanagement andmanagersofanorganizationdependstoalargeextentonthequalityofthedecisions theymakeonadailybasis.
*Correspondingauthor.
D.Stanujkicetal.
Decision-makinganddecisionsarethecoreofmanagerialactivities.Bearinginmind theglobalizationand,therefore,thedynamicsofbusiness doing,alloftheabove-stated havecausedbusinessandthedecision-makingprocesstobecomemoredemanding.Makingqualitydecisionsrequiresanevermoreextensivepreparation,whichalsoinvolvesthe considerationofthedifferentaspectsofadecision,forthe reasonofwhichthedecisionmakingprocessbecomesconsiderablyformalized.Thus,realproblemsandsituationsin reallifearecharacterizedbyalargenumberofmostlyconflictingcriteria,whosestrict optimizationisgenerallyimpossible.
Whenitisnecessarytomakeadecisiononchoosingoneofseveralpotentialsolutions toaproblem,itisdesirabletoapplyoneofthemodelsbasedonmultiple-criteriadecisionmakingmethods(MCDM).Thismostofteninvolvestheprocess ofselectingoneofseveral alternativesolutions,forwhichcertaingoalsareset.WhenMCDMisconcerned,Greco etal. (2010)pointoutthefactthatitisthestudyofthemethodsandproceduresaimed atmakingaproposalforsolutionsintermsofmultiple,oftenconflictingcriteria.Hwang andYoon(1981)statesthatMCDMisbasedonthetwobasicapproaches,i.e.onmultiple attributedecision-making(MADM),whichimpliesachoiceofcoursesinthepresence ofmultiple,andoftenconflictingcriteria,i.e.aselectionofthebestalternativefroma finitesetofpossiblealternatives.UnlikeMADM,inmultipleobjectivedecision-making (MODM),thebestalternativeisthatwhichisformedwithmultiplegoals,basedonthe continuousvariablesofthedecisionwithadditionalconstraints.
So,alltheproblemsoftodayare,ingeneral,multi-criterial,primarilybecauseproblemsaremainlyrelatedtotheachievementoftheobjectives relatedtoalargernumberof,usuallyconflicting,criteria,whichisagreatapproximationtorealtasksin decision-makingprocesses(Das etal.,2012;Zavadskas etal.,2014).TheincreasingapplicationoftheMCDMmethodtosolvingvariousproblemshas causedanexceptional growthofmulti-criteriadecision-makingasanimportantfieldofoperationalresearch, especiallysince1980(Aouni etal.,2018;Masri, etal.,2018;Wallenius etal.,2008; Dyer etal.,1992).
WithinMADM,someofthemethodsthathavebeenproposedare: WeightedSum Model(WSM)(Fishburn,1967);SimpleAdditiveWeighting(SAW)method(MacCrimon,1968),EliminationEtChoixTraduisantlaREalité(ELECTRE)method(Roy, 1968),DEcision-MAkingTrialandEvaluationLaboratory(DEMATEL)method(Gabus andFontela,1972),CompromiseProgramming(CP)method(Zeleny,1973),Simple MultiAttributeRatingTechnique(SMART)(Edwards,1977), AnalyticHierarchyProcess(AHP)method(Saaty,1978),TechniqueforOrderofPreferencebySimilarityto IdealSolution(TOPSIS)method(HwangandYoon,1981),PreferenceRankingOrganizationMethodforEnrichmentEvaluations(PROMETHEE)method(Brans,1982),MeasuringAttractivenessbyaCategoricalBasedEvaluationTechnique(MACBETH)(Banae CostaandVansnick,1994),ComplexProportionalAssessmentofalternatives(COPRAS) method(Zavadskas etal.,1994),AnalyticNetworkProcess(ANP)method(Saaty,1996), VlseKriterijumskaOptimizacijaikompromisnoResenje(VIKOR)(Opricovic,1998), Multi-ObjectiveOptimizationonbasisofRatioAnalysis(MOORA)method(Brauersand Zavadskas,2006),AdditiveRatioASsessment(ARAS)method (ZavadskasandTurskis,
ABipolarFuzzyExtensionoftheMULTIMOORAMethod 137 2010),Multi-ObjectiveOptimizationbyRatioAnalysisplustheFullMultiplicativeForm (MULTIMOORA)method(BrauersandZavadskas,2010a),andso on.Whilewithin theMODMmethodsthathavebeenproposedcanbestated:Dataenvelopmentanalysis (DEA)method(Charnes etal.,1978),LinearProgramming(LP)andNonlinearProgramming(NP)(LuenbergerandYe,1984),Multi-ObjectiveProgramming(MOP)technique (Charnes etal.,1989),Multi-ObjectiveLinearProgramming(EckerandKouada,1978), andsoon.
TheMULTIMOORAmethod(BrauersandZavadskas,2010b)isanimportantMCDM methodthathasbeenappliedsofartosolvethemostdiverseproblemsinthefieldofeconomics,management,etc.Basically,theMULTIMOORAmethod consistsofthewellknownMOORAmethod(BrauersandZavadskas,2006)andthemethodofmulti-object optimization(theFullMultiplicativeFormofMultipleObjectsmethod).Thus,Brauers andZavadskas(2010a)proposedtheupdatingoftheMOORAmethodbyaddingamultiobjectoptimizationmethodwhichinvolvesmaximizingandminimizingusefulmultiplicativefunctions(Lazauskas etal.,2015).
Asnotedabove,theMULTIMOORAmethodwasappliedinorderto solveavarietyof problems,suchas:usingMULTIMOORAforrankingandselectingthebestperformance appraisalmethod(Maghsoodi etal.,2018),projectcriticalpathselection(Dorfeshan etal., 2018),theselectionoftheoptimalminingmethod(Liang etal.,2018),pharmacological therapyselection(Eghbali-Zarch etal.,2018),ICThardwareselection(AdaliandIşik, 2017),industrialrobotselection(Karande etal.,2016),aCNCmachinetoolevaluation (Sahu etal.,2016),personnelselection(Karabasevic etal.,2015;Baležentis etal.,2012), theeconomy(BaležentisandZeng,2013;BrauersandZavadskas,2011a,2010b;Brauers andGinevičius,2010),andsoon.
However,mostdecisionsmadeintherealworldaremadeinanenvironmentinwhich goalsandconstraintscannotbepreciselyexpressedduetotheircomplexity;therefore,a problemcannotbedisplayedexactlyincrispnumbers(BellmanandZadeh,1970).For suchproblems,characterizedbyuncertaintyandindeterminacy,itismoreappropriateto usevaluesexpressedinintervalsinsteadofconcrete(crisp)values.Inthiscase,theexisting,ordinaryMCDMmethodsareexpandedbyusingtheextensionsbasedonfuzzysets (Zadeh,1965),intuitionisticfuzzysets(Atanassov,1986),andneutrosophicsets(Smarandache,1999).Accordingly,inordertoallowamuchwideruse oftheMULTIMOORA method,someextensionsoftheMULTIMOORAmethodhavebeenproposed,someof whichareasfollows:Brauers etal. (2011)proposedafuzzyextensionoftheMULTIMOORAmethod;BaležentisandZeng(2013)proposedanIVFNextensionoftheMULTIMOORAmethod;Baležentis etal. (2014)alsoproposedanIFNextensionoftheMULTIMOORAmethod;Stanujkic etal. (2015)proposedanextensionoftheMULTIMOORA methodbasedontheuseofinterval-valuedtriangularfuzzy numbers;Zavadskas etal. (2015)proposedanIVIF-basedextensionoftheMULTIMOORAmethod;Hafezalkotob etal. (2016)proposedanextensionoftheMULTIMOORAmethodbased ontheuseof intervalnumbers;Stanujkic etal. (2017a)proposedaneutrosophicextensionoftheMULTIMOORAmethod,andsoon.
Inadditiontotheaforementionedextensionsofthefuzzysettheory,Zhang(1994) introducedtheconceptofbipolarfuzzysetsandproposedtheusageofthetwomembership
D.Stanujkicetal.
functionsthatrepresentmembershiptoasetandmembership toacomplementaryset,thus providinganefficientapproachtosolvingawidelylargernumberofcomplexdecisionmakingproblems.
Despiteanadvantagethatcanbeachievedbyusingbipolarfuzzylogic,theyaresignificantlylessusedforsolvingMCDMproblemscomparedtootherfuzzylogicextensions. Thefollowingexamplescanbementionedassomeofthereally rareusagesofBFSfor solvingMCDMproblems:Alghamdi etal. (2018)andAkramandArshad(2018)proposedbipolarfuzzyextensionsofTOPSISandELECTREImethods;whileHan etal. (2018)provideacomprehensivemathematicalapproachbasedontheTOPSISmethod forimprovingtheaccuracyofbipolardisorderclinicaldiagnosis.
Itisalsoimportanttonotethatthesearecurrentresearches.Inaddition,thebipolarlogichasbeenconsiderablyusedintheneutrosophicset theory,whereUluçay etal. (2018),Pramanik etal. (018)andTian etal. (2016)canbecitedassomeofthecurrent researches.
Therefore,inordertoenableawideruseoftheMULTIMOORAmethodforsolving evenawiderrangeofproblems,abipolarextensionoftheMULTIMOORAmethodis proposedinthispaper.Accordingly,thepaperisstructuredasfollows:inSection1,the introductoryconsiderationsarepresented.InSection2,somebasicdefinitionsregardingbipolarfuzzysetsaregiven.InSection3,theordinaryMULTIMOORAmethodis presented,whereasinSection4,anextensionoftheMULTIMOORAmethodbasedon single-valuedbipolarfuzzynumbersisproposed.InSection5,anumericalexampleis demonstrated,andfinally,theconclusionsaregivenattheendofthepaper.
2.TheBasicElementsofaBipolarFuzzySet
Definition 1(Fuzzyset,Zadeh,1965).Let X beanonemptyset,withagenericelement in X denotedby x.Then,afuzzyset A in X isasetoforderedpairs:
A = x,µA(x) x ∈ X , (1)
wherethemembershipfunction µA(x) denotesthedegreeofthemembershipoftheelement x totheset A,and µA(x) ∈[0, 1]
Definition 2(Bipolarfuzzyset,Lee,2000).Let X beanonemptyset.Then,abipolar fuzzyset(BFS)isdefinedas:
A = x,µ+ A (x),νA (x) x ∈ X , (2) where:thepositivemembershipfunction µ+ A (x) denotesthesatisfactiondegreeofthe element x tothepropertycorrespondingtothebipolar-valuedfuzzyset,andthenegativemembershipfunction νA (x) denotesthedegreeofthesatisfactionoftheelement x tosomeimplicitcounter-propertycorrespondingtothebipolar-valuedfuzzyset,respectively; µ+ A (x) : X →[0, 1] and νA (x) : X →[−1, 0]
ABipolarFuzzyExtensionoftheMULTIMOORAMethod 139
Definition 3.Asingle-valuedbipolarfuzzynumber(SVBFN) a = a+,a isaspecialbipolarfuzzysetontherealnumberset R,whosepositivemembershipandnegative membershipfunctionareasfollows:
µ+(x) = 1 x = a+ , 0 otherwise, (3)
ν (x) = 1 x = a , 0 otherwise, (4) respectively.
Definition 4.Let a1 = a+ 1 ,a1 and a1 = a+ 2 ,a2 betwoSVBFNs,and λ> 0.Then, thebasicoperationsforthesenumbersaredefinedasshownbelow:
a1 + a2 = a+ 1 + a+ 2 a+ 1 a+ 2 , a1 a2 , (5) a1 · a2 = a+ 1 a+ 2 , a1 a2 a1 a2 , (6)
λa1 = 1 1 a+ 1 λ , a1 λ , (7) a λ 1 = a+ 1 λ , 1 1 a1 λ . (8)
Definition 5.Let a = a+,a beanSVBFN.Then,thescorefunction s(a) isasfollows:
sa = 1 + a+ + a 2 (9)
Definition 6.Let a1 and a2 betwoSVBFNs.Then, a1 >a2 if sa1 >sa2
Definition 7.Let a1 = a+ 1 ,a1 and a1 = a+ 2 ,a2 betwoSVBFNs.TheHamming distancebetween a1 and a2 isasfollows: dH (a1,a2) = 1 2 a+ 1 a+ 2 + a1 a2 (10)
Definition 8.Let aj = a+ j ,aj beacollectionofSVBFNs.Thebipolarweightedaverageoperator(Aw )ofthe n dimensionsisamappingasfollows:
Aw (a1,a2,...,an) = n j =1 wj aj = 1 n j =1 1 a+ j wj , 1 n j =1 1 aj wj , (11)
where: wj istheelement j oftheweightingvector, wj ∈[0, 1] and n j =1 wj = 1
140 D.Stanujkicetal. Definition 9.Let aj = a+ j ,aj beacollectionofSVBFNs.Thebipolarweightedgeometricoperator(Gw )ofthe n dimensionsisamapping Gw : Qn → Q asfollows: Gw (a1,a2,...,an) = n j =1 a wj j = n j =1 a+ j wj , n j =1 aj wj (12)
3.TheMULTIMOORAMethod
ComparedtotheotherMCDMmethods,theMULTIMOORAmethodis characteristic becauseitcombinesthreeapproaches,namely:theRatioSystem(RS)Approach,theReferencePoint(RP)ApproachandtheFullMultiplicativeForm (FMF)Approach,inorder toselectthemostappropriatealternative.
Inaddition,thismethoddoesnotcalculateanddoesnotusetheoverallsignificance forrankingalternativesandselectingthemostappropriateone.Insteadofusinganoverall parameterforrankingalternatives,thefinalrankingorder oftheconsideredalternatives, aswellastheselectionofthemostappropriatealternative,isbasedontheuseofthetheory ofdominance.
ForanMCDMproblemthatincludesthemalternativesthatshouldbeevaluatedon thebasisofthencriteria,thecomputationalprocedureoftheMULTIMOORAcanbe expressedasfollows:
Step1. Constructadecisionmatrixanddeterminetheweightsofcriteria.
Step2. Calculateanormalizeddecisionmatrix,asfollows: rij = xij n i=1 x 2 ij , (13)
where: rij denotesthenormalizedperformanceofthealternative i withrespecttothe criterion j ,and xij denotestheperformanceofthealternative i tothecriterion j
Step3. Calculatetheoverallsignificanceofeachalternative,asfollows: yi = j ∈ max
wj rij j ∈ min
wj rij , (14)
where: yi denotestheoverallimportanceofthealternative i, max and min denotethe setsofthebenefitcostcriteria,respectively.
Step4. Determinethereferencepoint,asfollows: r ∗ = r ∗ 1 ,r ∗ 2 ,...,r ∗ n = max i rij j ∈ max , min i rij j ∈ min . (15)
ABipolarFuzzyExtensionoftheMULTIMOORAMethod 141
Step5. Determinethemaximaldistancebetweeneachalternativeandthereferencepoint, asfollows: d max i = max j wj r ∗ j rij , (16) where: d max i denotesthemaximaldistanceofthealternative i tothereferencepoint. Step6. Determinetheoverallutilityofeachalternative,asfollows: ui = j ∈ max wj rij j ∈ min wj rij , (17) where: ui denotestheoverallutilityofthealternative i Inparticularcase,whenevaluationismadeonlyonthebasis ofbenefitcriteria,Eq.(17) isasfollows: ui = j ∈ max
wj rij (18) Step7. Determinethefinalrankingorderoftheconsideredalternativesandselectthe mostappropriateone.Inthisstep,theconsideredalternativesarerankedbased ontheir: –overallsignificance, –maximaldistancetothereferencepoint,and –overallutility.
Asaresultoftheserankings,thethreedifferentrankinglistsareformed,representing therankingsbasedontheRSapproach,theRPapproachandthe FMFapproachofthe MULTIMOORAmethod.
Thefinalrankingofthealternativesisbasedonthedominancetheory,i.e.thealternativewiththehighestnumberofappearancesinthefirstpositionsonallrankinglistsis thebest-rankedalternative.
4.AnExtensionoftheMULTIMOORAMethodBasedonSingle-ValuedBipolar FuzzyNumbers
ForanMCDMprobleminvolvingmalternativesandncriteriaandKdecision-makers, wherebytheperformancesofthealternativesareexpressed byusingSVBFNs,thecalculationprocedureoftheextendedMULTIMOORAmethodcanbeexpressedasfollows: Step1. Evaluatethealternativesinrelationtotheevaluationcriteria,anddothatforeach DM.Inthisstep,eachDMevaluatesthealternativesandformsanevaluationmatrix. Inordertoprovideaneasierevaluation,thefollowingLikertscale,showninTable1, isproposedforevaluatingalternativesinrelationtotheevaluationcriteria.
D.Stanujkicetal.
Table1 Nine-pointLikertscaleforexpressingdegreeofsatisfaction.
SatisfactionlevelNumericalvalue Neutral/withoutattitude0 Extremelylow1 Verylow2 Low3 Mediumlow4 Medium5 Mediumhigh6 High7 Veryhigh8 Extremelyhigh9 Absolute10
However,therespondentsshouldbeintroducedthatthevalueslistedinTable1are onlyapproximativeandthattheycanuseanyvaluefromtheinterval [0, 10] and [−10, 0].
Afterforminginitialdecision-makingmatrix,obtainedresponsesshouldbedividedby 10inordertotransformitintotheallowedinterval [−1, 1].Thisapproachforevaluating alternativesisproposedtoavoidtheuseofvectornormalizationprocedure,usedinthe ordinaryMULTIMOORAmethod.
Step2. Determinetheimportanceoftheevaluationcriteria,anddothatforeachDM.In thisstep,eachDMdeterminestheweightsofthecriteriabyusingoneofseveralexisting methodsfordeterminingtheweightsofcriteria.
Step3. Determinethegroupdecisionmatrix.Inordertotransformindividualintogroup preferences,individualevaluationmatricesaretransformedintogrouponebyapplying Eq.(11).
Step4. Determinethegroupweightsofthecriteria.Inordertotransformindividualinto grouppreferenceswithrespecttotheweightsofcriteria,thegroupweightsofcriteriacan bedeterminedasfollows:
(19)
where: wj denotestheweightofthecriterion j ,and wk j denotestheweightofthecriterion j obtainedfromtheDM k.
Aftercalculatingthegroupevaluationmatrixandthegroup weightsofthecriteria, allthenecessaryprerequisitesforapplyingallthethreeapproachesintegratedinthe MULTIMOORAmethodareobtained.BasedontheapproachproposedbyStanujkic et al. (2017b),theremainderstepsoftheproposedapproachareas follows:
Step5. DeterminethesignificanceoftheevaluatedalternativesbasedontheRSapproach Thisstepcanbeexplainedthroughthefollowingsub-steps:
ABipolarFuzzyExtensionoftheMULTIMOORAMethod 143
Step5.1. Determinetheimpactofthebenefitandcostcriteriatotheimportanceofeach alternative,asfollows:
Y + i = 1 n j ∈ max
(1 rij )wj , 1 n j ∈ max
1 ( rij ) wj , (20) Yi = 1 n j ∈ min
(1 rij )wj , 1 n j ∈ min
1 ( rij ) wj , (21)
where: Y + i and Yi denotetheimportanceofthealternative i obtainedonthebasisofthe benefitandcostcriteria,respectively; Y + i and Yi areSVBFNs.
Itisevidentthat Aw operatorisusedtocalculatetheimpactofthebenefitandcost criteria.
Step.5.2. Transform Y + i and Yi intocrispvaluesbyusingtheScoreFunction,asfollows: y+ i = s Y + i , (22) yi = s Yi (23)
Step5.3. Calculatetheoverallimportanceforeachalternative,asfollows: yi = y+ i yi (24)
Step6. DeterminethesignificanceoftheevaluatedalternativesbasedontheRPapproach Thisstepcanbeexplainedthroughthefollowingsub-steps:
Step6.1. Determinethereferencepoint.Thecoordinatesonthebipolarfuzzyreference point r ∗ ={r ∗ 1 ,r ∗ 2 ,...,r ∗ n } canbedeterminedasfollows: r ∗ = max i rij , min i rij j ∈ max , min i rij , max i rij j ∈ min (25) where: r ∗ j denotesthecoordinate j ofthereferencepoint.
Step6.2. Determinethemaximumdistancefromeachalternativetoall thecoordinatesof thereferencepoint.Themaximumdistanceofeachalternativetothereferencepointcan bedeterminedasfollows: d max ij = dmax rij ,r ∗ j wj , (26) where d max ij denotesthemaximumdistanceofthealternative i tothecriterion j determinedbyEq.(10).
Step6.3. Determinethemaximumdistanceofeachalternative,asfollows: d max i = max j d max ij , (27) where d max i denotesthemaximumdistanceofthealternative i
144
D.Stanujkicetal.
Step7. Determinethesignificanceoftheevaluatedalternativesbasedon theFMF. Thisstepcanbeexplainedthroughthefollowingsub-steps:
Step7.1. Calculatetheutilityobtainedbasedonthebenefit U + i andcost Ui criteria,for eachalternative,asfollows: U + i = n j ∈ max
(rij )wj , n j ∈ max
( rij )wj , (28) Ui = n j ∈ min
(rij )wj , n j ∈ min
where U + i and Ui areSVBFNs.
( rij )wj , (29)
Step7.2. Transform U + i and Ui intocrispvaluesbyusingtheScoreFunction,asfollows: u+ i = s U + i , (30) ui = s Ui . (31)
Step7.3. Determinetheoverallutilityforeachalternative,asfollows: ui = u+ i ui . (32)
Inthecasewhenevaluationismadeonlyonthebasisofbenefit criteria,Eq.(32)isas follows: ui = u+ i . (33)
Step8. Determinethefinalrankingorderofthealternatives.Thefinalrankingorderof thealternativescanbedeterminedasinthecaseoftheordinaryMULTIMOORAmethod, i.e.basedonthedominancetheory(BrauersandZavadskas,2011b).
Inthisstage,thealternativesarerankedbasedontheiroverallimportance,maximum distancetothereferencepointandoverallutility.Asaresultofthat,threerankinglistsare formed.
Basedontheserankinglists,thefinalrankinglistofthealternativesisformedonthe basisofthetheoryofdominance,i.e.thealternativewiththelargestnumberofappearancesonthefirstpositioninthethreerankinglistsisthemostacceptable.
5.ANumericalExample
Inthissection,anumericalexampleofpurchasingrentalspaceisconsideredinorderto explaintheproposedapproachindetail.
ABipolarFuzzyExtensionoftheMULTIMOORAMethod 145
Table2
TheratingsobtainedfromthefirstofthethreeDMs.
C1 C2 C3 C4 C5 a+ a a+ a a+ a a+ a a+ a
A1 7 27 35 17 58 1
A2 4 15 24 24 67 1
A3 7 13 1202 17 2 A4 9 14 1303 16 1
Table3
TheratingsobtainedfromthefirstofthethreeDMs,intheformofSVBFNs.
C1 C2 C3 C4 C5
A1 0 70, 0 20 0 70, 0 30 0 50, 0 10 0 70, 0 50 0 80, 0 10
A2 0 40, 0 10 0 50, 0 20 0 40, 0 20 0 40, 0 60 0 70, 0 10
A3 0 70, 0 10 0 30, 0 10 0 20, 0 00 0 20, 0 10 0 70, 0 20
A4 0 90, 0 10 0 40, 0 10 0 30, 0 00 0 30, 0 10 0 60, 0 10
Table4
TheratingsobtainedfromthesecondofthethreeDMs,intheformofSVBFNs.
C1 C2 C3 C4 C5
A1 0 70, 0 20 0 70, 0 50 0 40, 0 20 0 70, 0 50 0 80, 0 10 A2 0 60, 0 10 0 40, 0 60 0 40, 0 20 0 40, 0 60 0 80, 0 10
A3 0.80, 0.10 0.20, 0.10 0.20, 0.10 0.20, 0.10 0.70, 0.10
A4 0 90, 0 10 0 30, 0 10 0 30, 0 10 0 30, 0 10 0 60, 0 10
Supposethatacompanyisplanningtostartitssalesbusinessinanewlocation,and thereforeislookingforanewsalesbuilding.Aftertheinitialconsiderationoftheavailable alternatives,fouralternativeshavebeenidentifiedassuitable.Forthisreason,ateamof threedecision-makers(DMs)wasformedwiththeaimofevaluatingsuitablealternatives basedonthefollowingcriteria: C1 –Rentalspacequality; C2 –Rentalspaceadequacy; C3 –Locationquality; C4 –Locationdistancefromthecitycentre,and C5 –Rentalprice.
Aspreviouslyreasoned,inthisevaluationtheratingsofthealternativesinrelationto thecriteriaareexpressedbyusingSVBFNs.
TheratingsobtainedfromthefirstofthethreeDMsareshowninTable2,asthepoints oftheLikertscale,whereasinTable3,theyareshownintheformofSVBFNs.
TheratingsobtainedfromthesecondandthethirdofthethreeDMsareaccountedfor inTable4andTable5.
Thegroupdecisionmatrix,calculatedbyapplyingEq.(11), ispresentedinTable6.
D.Stanujkicetal.
Table5
TheratingsobtainedfromthethirdofthethreeDMs,intheformofSVBFNs.
C1 C2 C3 C4 C5
A1 0 60, 0 10 0 90, 0 20 1 00, 0 00 1 00, 0 00 0 80, 0 10
A2 0 40, 0 60 0 40, 0 60 1 00, 0 40 1 00, 0 00 0 80, 0 10
A3 0 20, 0 10 0 90, 0 40 0 80, 0 30 0 70, 0 10 0 70, 0 10
A4 0 30, 0 10 1 00, 0 30 0 80, 0 20 0 80, 0 10 0 60, 0 10
Table6
Thegroupdecision-makingmatrix.
C1 C2 C3 C4 C5
A1 0 67, 0 16 0 79, 0 32 1 00, 0 00 1 00, 0 00 0 80, 0 10
A2 0 47, 0 18 0 43, 0 42 1 00, 0 26 1 00, 0 00 0 77, 0 10
A3 0 64, 0 10 0 61, 0 16 0 49, 0 00 0 41, 0 10 0 70, 0 13
A4 0 81, 0 10 1 00, 0 15 0 53, 0 00 0 53, 0 10 0 60, 0 10
Table7
Theweightsofthecriteriaobtainedfromthefirstofthethree DMs.
sj kj qj wj
C1 110.19 C2 1.20.801.250.23 C3 0.91.101.140.21 C4 0.71.300.870.16 C5 1.20.801.090.20 5.005.351.00
Table8
Thegroupcriteriaweights.
w1 j w2 j w3 j wj C1 0.190.170.190.18 C2 0.230.240.230.24 C3 0.210.220.210.21 C4 0.160.170.160.16 C5 0.200.210.200.21 1.00
TheweightsobtainedfromthefirstofthethreeDMsbyapplyingthePIPRECIA method(Stanujkic etal.,2017b)areaccountedforinTable7,whilethegroupweights ofthecriteria,calculatedbyapplyingEq.(19),areshowninTable8.
OnthebasisoftheratingsfromTable6andtheweightsfromTable8,theoverallsignificance,themaximumdistancetothereferencepointandtheoverallutilityarecalculated foreachalternativeinthenextstep.
Theoverallsignificances,accountedforinTable9,arecalculatedbyapplying Eqs.(20)–(24).
Table9
Theoverallsignificancesoftheconsideredalternatives.
Y + i Yi y+ i yi yi Rank
A1 1 00, 0 11 1 00, 0 02 0.940.99-0.053
A2 1 00, 0 20 1 00, 0 02 0.900.99-0.094 A3 0 42, 0 06 0 30, 0 05 0.680.630.052
A4 1 00, 0 06 0 29, 0 04 0.970.620.351
Table10 Thereferencepoints.
max min r+ r r+ r r∗ 1.00 0.200.29 0.02
Table11
Theratingsofthealternativesobtainedbasedonthereferencepointapproach.
C1 C2 C3 C4 C5 d max i Rank
A1 0.080.160.130.290.100.084 A2 0.170.280.000.290.090.001 A3 0.130.330.380.050.060.053 A4 0.040.140.360.110.000.001
Table12
Theoverallutilityoftheconsideredalternatives.
U + i Ui u+ i ui ui Rank
A1 1 00, 0 11 1 00, 0 02 0.940.99 0.053
A2 1 00, 0 20 1 00, 0 02 0.900.99 0.094 A3 0 42, 0 06 0 30, 0 05 0.680.630.052 A4 1 00, 0 06 0 29, 0 04 0.970.620.351
Afterthat,thereferencepointshowninTable10isdeterminedbyapplyingEq.(25).
ThemaximumdistancestothereferencepointaccountedforinTable11aredeterminedbyapplyingEq.(26)andEq.(27).
TheoverallutilityshowninTable12iscalculatedbyapplyingEqs.(28)–(32).
Finally,onthebasisoftherankingordersshowninTables9, 11and12,themost appropriatealternativeisdeterminedbymeansofthetheoryofdominance,asisshown inTable13.
AscanbeseenfromTable12,themostappropriatealternativeisthealternative denotedas A4
D.Stanujkicetal.
Table13
Thefinalrankingorderoftheconsideredalternatives.
RSRPFMP Finalrank
A1 3433 A2 4144 A3 2322 A4 1111
6.Conclusions
Thebipolarfuzzysetsintroducedtwomembershipfunctions,namelythemembership functiontoasetandthemembershipfunctiontoacomplementaryset.
Ontheotherhand,theMULTIMOORAmethodisanefficientandalreadyproven multiple-criteriadecision-makingmethod,whichhasbeen usedforsolvinganumberof differentdecision-makingproblemssofar.
Therefore,anextensionoftheMULTIMOORAmethodenablingtheuseofsinglevaluedbipolarfuzzynumbersisproposedinthisarticle.Theusabilityandefficiencyof theproposedextensionissuccessfullydemonstratedonthe exampleoftheproblemofthe bestlocationselection.
Intheliterature,numerousextensionsoftheMULTIMOORAmethodshavebeenproposedwiththeaimtoadaptitfortheuseofgreysystemtheory,fuzzysettheory,aswell asvariousextensionsoffuzzysettheory.Someextensionsthatenabletheuseofneutrosophicsetsarealsoproposed.Thementionedextensionsaim toexploitthespecificities ofparticularsetsforsolvingcertaintypesofdecision-makingproblems,andthusenable moreefficientdecisionmaking.
Becauseofthespecificitythatbipolarfuzzysetsprovide,theproposedexpanded MULTIMOORAmethodcanbeexpectedtobeacceptableforsolvingaparticularclass ofcomplexdecision-makingproblems.
References
Adali,E.A.,Işik,A.T.(2017).Themulti-objectivedecisionmakingmethodsbasedonMULTIMOORAand MOOSRAforthelaptopselectionproblem. JournalofIndustrialEngineeringInternational,13(2),229–237.
Akram,M.,Arshad,M.(2018).AnoveltrapezoidalbipolarfuzzyTOPSISmethodforgroupdecision-making. GroupDecisionandNegotiation,1–20.
Alghamdi,M.A.,Alshehri,N.O.,Akram,M.(2018).Multi-criteriadecision-makingmethodsinbipolarfuzzy environment. InternationalJournalofFuzzySystems,20(6),2057–2064.
Aouni,B.,Doumpos,M.,Pérez-Gladish,B.,Steuer,R.E.(2018).Ontheincreasingimportanceofmultiple criteriadecisionaidmethodsforportfolioselection. JournaloftheOperationalResearchSociety,69(10), 1525–1542.
Atanassov,K.T.(1986).Intuitionisticfuzzysets. FuzzySetsandSystems,20(1),87–96.
Baležentis,T.,Zeng,S.(2013).Groupmulti-criteriadecisionmakingbaseduponinterval-valuedfuzzynumbers: anextensionoftheMULTIMOORAmethod. ExpertSystemswithApplications,40(2),543–550.
Baležentis,A.,Baležentis,T.,Brauers,W.K.(2012).MULTIMOORA-FG:amulti-objectivedecisionmaking methodforlinguisticreasoningwithanapplicationtopersonnelselection. Informatica,23(2),173–190.
ABipolarFuzzyExtensionoftheMULTIMOORAMethod 149
Baležentis,T.,Zeng,S.,Balezentis,A.(2014),MULTIMOORA-IFN:aMCDMmethodbasedonintuitionistic fuzzynumberforperformancemanagement, EconomicComputationandEconomicCyberneticsStudiesand Research,48(4),85–102.
BanaeCosta,C.A.,Vansnick,J.C.(1994).MACBETH–aninteractivepathtowardstheconstructionofcardinal valuefunctions. InternationalTransactionsinOperationalResearch,1(4),489–500.
Bellman,R.E.,Zadeh,L.A.(1970).Decision-makinginafuzzyenvironment. ManagementScience,17(4),141–164.
Brans,J.P.,(1982).Língénieriedeladécision.Elaborationdínstrumentsdáideàladécision.Méthode PROMETHEE.In:Nadeau,R.,Landry,M.(Eds.), L´aidealaDécision:Nature,InstrumentsetPerspectivesd´avenir.Pressesdel´UniversitéLaval,Québec,Canada,pp.183-214.
Brauers,W.K.M.,Ginevičius,R.(2010).TheeconomyoftheBelgianregionstestedwithMULTIMOORA. JournalofBusinessEconomicsandManagement,11(2),173–209.
Brauers,W.K.,Zavadskas,E.K.(2006).TheMOORAmethodand itsapplicationtoprivatizationinatransition economy. ControlandCybernetics,35(2),445–469.
Brauers,W.K.M.,Zavadskas,E.K.(2010a).ProjectmanagementbyMULTIMOORAasaninstrumentfortransitioneconomies. TechnologicalandEconomicDevelopmentofEconomy,16(1),5–24.
Brauers,W.K.M.,Zavadskas,E.K.(2010b).RobustnessintheMULTIMOORAmodel:theexampleofTanzania. TransformationsinBusinessandEconomics,9(3),67–83.
Brauers,W.K.M.,Zavadskas,E.K.(2011a).Fromacentrally plannedeconomytomultiobjectiveoptimization inanenlargedprojectmanagement:thecaseofChina. EconomicComputationandEconomicCybernetics StudiesandResearch,45(1),167–188.
Brauers,W.K.M.,Zavadskas,E.K.(2011b).Multimooraoptimizationusedtodecideonabankloantobuy property. TechnologicalandEconomicDevelopmentofEconomy,17(1),174–188.
Brauers,W.K.,Balezentis,A.,Balezentis,T.(2011).MULTIMOORAfortheEUMemberStatesupdatedwith fuzzynumbertheory. TechnologicalandEconomicDevelopmentofEconomy,17(2),259–290.
Charnes,A.,Cooper,W.W.,Rhodes,E.(1978).Measuringthe efficiencyofdecisionmakingunits. European journalofoperationalresearch,2(6),429–444.
Charnes,A.,Cooper,W.W.,Wei,Q.L.,Huang,Z.M.(1989).Coneratiodataenvelopmentanalysisandmultiobjectiveprogramming. InternationalJournalofSystemsScience,20(7),1099–1118.
Das,M.C.,Sarkar,B.,Ray,S.(2012).Decisionmakingunder conflictingenvironment:anewMCDMmethod. InternationalJournalofAppliedDecisionSciences,5(2),142–162.
Dorfeshan,Y.,Mousavi,S.M.,Mohagheghi,V.,Vahdani,B.(2018).Selectingproject-criticalpathbyanewintervaltype-2fuzzydecisionmethodologybasedonMULTIMOORA,MOOSRAandTPOPmethods. ComputersandIndustrialEngineering,120,160–178.
Dyer,J.S.,Fishburn,P.C.,Steuer,R.E.,Wallenius,J.,Zionts,S.(1992).Multiplecriteriadecisionmaking,multiattributeutilitytheory:thenexttenyears. ManagementScience,38(5),645–654.
Ecker,J.G.,Kouada,I.A.(1978).Findingallefficientextremepointsformultipleobjectivelinearprograms. MathematicalProgramming,14(1),249–261.
Edwards,W.(1977).12useofmultiattributeutilitymeasurementforsocialdecisionmaking. Conflicting,247.
Eghbali-Zarch,M.,Tavakkoli-Moghaddam,R.,Esfahanian, F.,Sepehri,M.M.,Azaron,A.(2018).Pharmacologicaltherapyselectionoftype2diabetesbasedontheSWARAandmodifiedMULTIMOORAmethods underafuzzyenvironment. ArtificialIntelligenceinMedicine,87,20–33.
Fishburn,P.C.(1967).Lettertotheeditor–additiveutilitieswithincompleteproductsets:applicationtoprioritiesandassignments. OperationsResearch,15(3),537–542.
Gabus,A.,Fontela,E.(1972). WorldProblems,anInvitationtoFurtherThoughtwithinthe FrameworkofDEMATEL.BattelleGenevaResearchCenter,Geneva,Switzerland,1–8.
Greco,S.,Slowiński,R.,Figueira,J.R.,Mousseau,V.(2010).Robustordinalregression.In: TrendsinMultiple CriteriaDecisionAnalysis.Springer,Boston,MA,pp.241–283.
Hafezalkotob,A.,Hafezalkotob,A.,Sayadi,M.K.(2016).ExtensionofMULTIMOORAmethodwithinterval numbers:anapplicationinmaterialsselection. AppliedMathematicalModelling,40(2),1372–1386.
Han,Y.,Lu,Z.,Du,Z.,Luo,Q.,Chen,S.(2018).AYinYangbipolarfuzzycognitiveTOPSISmethodtobipolar disorderdiagnosis. ComputerMethodsandProgramsinBiomedicine,158,1–10.
Hwang,C.L.,Yoon,K.(1981).Methodsformultipleattributedecisionmaking.In MultipleAttributeDecision Making.Springer,Berlin,Heidelberg,58–191.
Karabasevic,D.,Stanujkic,D.,Urosevic,S.,Maksimovic, M.(2015).Selectionofcandidatesinthemining industrybasedontheapplicationoftheSWARAandtheMULTIMOORAmethods. ActaMontanistica Slovaca,20(2),116–124.
150 D.Stanujkicetal.
Karande,P.,Zavadskas,E.,Chakraborty,S.(2016).Astudy ontherankingperformanceofsomeMCDMmethodsforindustrialrobotselectionproblems. InternationalJournalofIndustrialEngineeringComputations, 7(3),399–422.
Lazauskas,M.,Kutut,V.,Zavadskas,E.K.(2015).Multicriteriaassessmentofunfinishedconstructionprojects. Gradevinar,67(04.),319–328.
Lee,K.M.(2000).Bipolar-valuedfuzzysetsandtheirbasic operations.In: ProceedingInternationalConference, Bangkok,Thailand,pp.307–317.
Liang,W.,Zhao,G.,Hong,C.(2018).Selectingtheoptimalminingmethodwithextendedmulti-objectiveoptimizationbyratioanalysisplusthefullmultiplicativeform(MULTIMOORA)approach. NeuralComputing andApplications,1–16.
Luenberger,D.G.,Ye,Y.(1984). LinearandNonlinearProgramming,Vol.2.Addison-Wesley,Reading,MA. MacCrimon,K.R.(1968). DecisionMarkingAmongMultiple-AttributeAlternatives: aSurveyandConsolidated Approach.RANDMemorandum,RM-4823-ARPA.TheRandCorporation,SantaMonica,CA.
Maghsoodi,A.I.,Abouhamzeh,G.,Khalilzadeh,M.,Zavadskas,E.K.(2018).Rankingandselectingthebest performanceappraisalmethodusingtheMULTIMOORAapproachintegratedShannon’sentropy. Frontiers ofBusinessResearchinChina,12(1),2.
Masri,H.,Pérez-Gladish,B.,Zopounidis,C.(2018). FinancialDecisionAidUsingMultipleCriteria:Recent ModelsandApplications.Springer.
Opricovic,S.(1998).Multicriteriaoptimizationofcivil engineeringsystems. FacultyofCivilEngineering,Belgrade,2(1),5–21.
Pramanik,S.,Dalapati,S.,Roy,T.K.(2018).Neutrosophic multi-attributegroupdecisionmakingstrategyfor logisticscenterlocationselection. NeutrosophicOperationalResearch,3,13–32.
Roy,B.(1968).Classementetchoixenprésencedepointsdevuemultiples. Revuefraniaised’informatiqueet derechercheopérationnelle,2(8),57–75.
Saaty,T.L.(1978).Modelingunstructureddecisionproblems–thetheoryofanalyticalhierarchies. Mathematics andComputersinSimulation,20(3),147–158.
Saaty,T.L.(1996). TheANPforDecisionMakingwithDependenceandFeedback.RWSPublications,USA. Sahu,A.K.,Sahu,N.K.,Sahu,A.K.(2016).ApplicationofmodifiedMULTI-MOORAforCNCmachinetool evaluationinIVGTFNSenvironment:anempiricalstudy. InternationalJournalofComputerAidedEngineeringandTechnology,8(3),234–259.
Smarandache,F.(1999). AUnifyingFieldinLogics.Neutrosophy:NeutrosophicProbability,SetandLogic AmericanResearchPress,Rehoboth.
Stanujkic,D.,Zavadskas,E.K.,Brauers,W.K.,Karabasevic,D.(2015).AnextensionoftheMULTIMOORA methodforsolvingcomplexdecision-makingproblemsbased ontheuseofinterval-valuedtriangularfuzzy numbers. TransformationsinBusinessandEconomics,14(2B),355–377.
Stanujkic,D.,Zavadskas,E.K.,Karabasevic,D.,Smarandache,F.,Turskis,Z.(2017a).Theuseofthepivot pairwiserelativecriteriaimportanceassessmentmethodfordeterminingtheweightsofcriteria. Romanian JournalofEconomicForecasting,20(4),116–133.
Stanujkic,D.,Zavadskas,E.K.,Smarandache,F.,Brauers, W.K.,Karabasevic,D.(2017b).AneutrosophicextensionoftheMULTIMOORAmethod. Informatica,28(1),181–192.
Tian,Z.P.,Zhang,H.Y.,Wang,J.,Wang,J.Q.,Chen,X.H.(2016).Multi-criteriadecision-makingmethodbased onacross-entropywithintervalneutrosophicsets. InternationalJournalofSystemsScience,47(15),3598–3608.
Uluçay,V.,Deli,I.,Şahin,M.(2018).Similaritymeasures ofbipolarneutrosophicsetsandtheirapplicationto multiplecriteriadecisionmaking. NeuralComputingandApplications,29(3),739–748.
Wallenius,J.,Dyer,J.S.,Fishburn,P.C.,Steuer,R.E.,Zionts,S.,Deb,K.(2008).Multiplecriteriadecision making,multiattributeutilitytheory:Recentaccomplishmentsandwhatliesahead. ManagementScience, 54(7),1336–1349.
Zadeh,L.A.(1965).Fuzzysets. InformationandControl,8(3),338–353.
Zavadskas,E.K.,Turskis,Z.(2010).Anewadditiveratioassessment(ARAS)methodinmulticriteriadecisionmaking. TechnologicalandEconomicDevelopmentofEconomy,16(2),159–172.
Zavadskas,E.K.,Kaklauskas,A.,Sarka,V.(1994).Thenewmethodofmulticriteriacomplexproportionalassessmentofprojects. TechnologicalandEconomicDevelopmentofEconomy,1(3),131–139.
Zavadskas,E.K.,Turskis,Z.,Kildien˙e,S.(2014).StateofartsurveysofoverviewsonMCDM/MADMmethods. TechnologicalandEconomicDevelopmentofEconomy,20(1),165–179.
ABipolarFuzzyExtensionoftheMULTIMOORAMethod 151
Zavadskas,E.K.,Antucheviciene,J.,RazaviHajiagha,S.H.,Hashemi,S.S.(2015).Theinterval-valuedintuitionisticfuzzyMULTIMOORAmethodforgroupdecisionmakinginengineering. MathematicalProblems inEngineering.Art.No.560690.
Zeleny,M.(1973).Compromiseprogramming.In:Cochrane,J.L.,Zeleny,M.(Eds.). MultipleCriteriaDecision Making.UniversityofSouthCarolinaPress,Columbia,SC,pp.262–301.
Zhang,W.R.(1994).Bipolarfuzzysetsandrelations:acomputationalframeworkforcognitivemodelingand multiagentdecisionanalysis.In ProceedingsofIndustrialFuzzyControlandIntelligentSystemsConference, December18–21,1994,SanAntonio,USA,pp.305–309.
D.Stanujkic isanassociateprofessorofinformationtechnologyattheTechnicalFaculty inBor,UniversityofBelgrade.HehasreceivedhisMScdegreeininformationscienceand PhDinorganizationalsciencesfromtheFacultyofOrganizationalSciences,University ofBelgrade.Hiscurrentresearchisfocusedondecision-makingtheory,expertsystems andintelligentdecisionsupportsystems.
D.Karabasevic isanassistantprofessorattheFacultyofAppliedManagement,EconomicsandFinance,UniversityBusinessAcademyinNoviSad.Heobtainedhisdegrees atallthelevelsofstudies(BScappl.ineconomics,BScineconomics,academicspecializationinthemanagementofbusinessinformationsystemsandPhDinmanagementand business)attheFacultyofManagementinZajecar,JohnNaisbittUniversity,Belgrade. Hiscurrentresearchisfocusedonthehumanresourcemanagement,managementand decision-makingtheory.
E.K.Zavadskas isaprofessoroftheDepartmentofConstructionManagement andReal Estate,directorofInstituteofSustainableConstruction,andChiefResearchFellowof LaboratoryofOperationalResearchatVilniusGediminasTechnicalUniversity,Vilnius, Lithuania.HehasaPhDinbuildingstructures(1973)andDrSc(1987)inbuildingtechnologyandmanagement.HeisamemberoftheLithuanianandseveralforeignAcademies ofSciences.HeisdoctorhonoriscausaatPoznan,SaintPetersburg,andKievuniversities.Heistheeditorinchiefandamemberofeditorialboardsofanumberofresearch journals.Heisanauthorandcoauthorofmorethan400papers andanumberofmonographs.Researchinterestsare:buildingtechnologyandmanagement,decision-making theory,automationindesignanddecisionsupportsystems.
D.Stanujkicetal.
F.Smarandache isaprofessorofmathematicsattheUniversityofNewMexico,USA. Hehaspublishedmanypapersandbooksonneutrosophicsetandlogicandtheirapplicationsandhaspresentedtomanyinternationalconferences. HegothisMScinmathematics andcomputersciencefromtheUniversityofCraiova,Romania,PhDfromtheStateUniversityofKishinev,andpost-doctoralinappliedmathematicsfromOkayamaUniversity ofSciences,Japan.
W.K.M.Brauers,doctorhonoriscausaVilniusGediminasTechnicalUniversity,was graduatedas:PhDineconomics(un.ofLeuven),MasterofArts(ineconomics)of ColumbiaUniversity(NewYork),masterineconomics,masterinmanagementandfinancialsciences,masterinpoliticalanddiplomaticsciencesandbachelorinphilosophy (allintheUniversityofLeuven).HeisprofessorordinariusattheFacultyofApplied EconomicsoftheUniversityofAntwerp,honoraryprofessor attheUniversityofLeuven, theBelgianWarCollege,theSchoolofMilitaryAdministratorsandtheAntwerpBusinessSchool.Hisscientificpublicationsconsistofeighteenbooksandseveralhundredsof articlesandreportsinEnglish,DutchandFrench.