Delft University of Technology
Piecewise linear value functions for multi-criteria decision-making
Rezaei, Jafar
DOI
10.1016/j.eswa.2018.01.004
Publication date
2018
Document Version
Final published version
Published in
Expert Systems with Applications
Citation (APA)
Rezaei, J. (2018). Piecewise linear value functions for multi-criteria decision-making. Expert Systems with
Applications, 98, 43-56. https://doi.org/10.1016/j.eswa.2018.01.004
Important note
To cite this publication, please use the final published version (if applicable).
Please check the document version above.
Copyright
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons. Takedown policy
Please contact us and provide details if you believe this document breaches copyrights. We will remove access to the work immediately and investigate your claim.
This work is downloaded from Delft University of Technology.
Green Open Access added to TU Delft Institutional Repository
‘You share, we take care!’ – Taverne project
https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher
is the copyright holder of this work and the author uses the
Dutch legislation to make this work public.
ContentslistsavailableatScienceDirect
Expert
Systems
With
Applications
journalhomepage:www.elsevier.com/locate/eswa
Piecewise
linear
value
functions
for
multi-criteria
decision-making
Jafar
Rezaei
Faculty of Technology Policy and Management, Delft University of Technology, Delft, The Netherlands
a
r
t
i
c
l
e
i
n
f
o
Article history: Received 24 August 2017 Revised 1 January 2018 Accepted 2 January 2018 Available online 3 January 2018 Keywords: Multi-criteria decision-making MCDM Decision criteria Value function Monotonicity
a
b
s
t
r
a
c
t
Multi-criteriadecision-making(MCDM)concernsselecting,rankingorsortingasetofalternativeswhich areevaluatedwithrespecttoanumber ofcriteria.ThereareseveralMCDMmethods,thetwocore el-ementsofwhichare(i)evaluatingtheperformance ofthealternativeswithrespecttothecriteria,(ii) findingtheimportance(weight)ofthecriteria.Thereareseveralmethodstofindtheweightsofthe cri-teria,however,whenitcomestothealternativemeasureswithrespecttothecriteria,usuallytheexisting MCDMmethodsusesimplemonotoniclinearvaluefunctions.Usuallyanincreasingordecreasinglinear functionisassumedbetweenacriterionlevel(overitsentirerange)anditsvalue.Thisassumption, how-ever,mightleadtoimproperresults.Thisstudyproposesafamilyofpiecewisevaluefunctionswhichcan beusedfordifferentdecisioncriteriafordifferentdecisionproblems.Severalreal-worldexamplesfrom existingliteratureareprovidedtoillustratetheapplicabilityoftheproposedvaluefunctions.Anumerical exampleofsupplierselection(includingacomparisonbetweensimplemonotoniclinearvaluefunctions, piecewiselinearvaluefunctions,andexponentialvaluefunctions)showshowconsideringpropervalue functionscouldaffectthefinalresultsofanMCDMproblem.
© 2018ElsevierLtd.Allrightsreserved.
1. Introduction
Decisiontheoryisprimarilyconcernedwithidentifyingthebest
decision. In many real-world situations the decision is to select
the bestalternative(s) fromamong aset ofalternatives
consider-ing a set of criteria. This subdivision of decision-making, which
has gained enormousattention, due to its practical value, in the
pastrecentiscalledmulti-criteriadecision-making(MCDM).More
precisely, MCDMconcerns problemsin whichthe decision-maker
faces m alternatives (a1,a2,…, am), which should be evaluated
withrespect toncriteria(c1,c2,…, cn), inorderto findthe best
alternative(s), rankorsortthem.Inmostcases,an additive value
functionisusedtofindtheoverallvalueofalternativei,Ui,as
fol-lows: Ui= n j=1 wjui j, (1)
where uij is the value of alternative i withrespect to criterion j,
andwjshowstheimportance(weight)ofcriterionj.Insome
prob-lems,thedecision-makerisabletofinduijfromexternalsourcesas
objectivemeasures,insome otherproblems,uij reflectsa
qualita-tiveevaluationprovidedbythedecision-maker(s),expertsorusers
assubjectivemeasures.Priceofacarisanobjectivecriterionwhile
comfortofacarisasubjectiveone.Forobjectivecriteria,we
usu-E-mail address: j.rezaei@tudelft.nl
allyuse physicalquantities, for instance,‘InternationalSystem of
Units’(SI),whileforsubjectivecriteria,wedonothavesuch
stan-dards,which iswhywemostlyusepairwise comparison,
linguis-ticvariables,or Likertscales inorder toevaluate thealternatives
withregard tosuch criteria. Inorder tofindthe weights, wj,the
decision-makermight use different tools and methods, from the
simplestway,whichisassigningweightstothecriteriaintuitively,
touse simple methodslike SMART (simple multi-attribute rating
technique)(Edwards,1977),tomorestructuredmethodslike
mul-tipleattributeutility theory(MAUT)(Keeney&Raiffa, 1976),
ana-lytichierarchyprocess(AHP)(Saaty,1977),andbestworstmethod
(BWM) (Rezaei, 2015, 2016). While these methods are usually
called‘multi attribute utility and value theories’ (Carrico, Hogan,
Dyson,& Athanassopoulos, 1997),there isanother classof
meth-ods, called outranking methods, like ELECTRE (ELimination and
ChoiceExpressingREality)family (Roy, 1968), PROMETHEE
meth-ods (Brans, Mareschal, & Vincke, 1984) which do not necessarily
need the weights to select, rank or sort the alternatives. What,
however,isincommoninthesemethodsisthewaytheyconsider
thenatureofthe criteria.Thatisto say,inthe currentliterature,
one of the commonassumptions aboutthe criteria (most of the
timeitisnotexplicitlymentionedintheliterature),is
monotonic-ity.
Definition1 (Keeney&Raiffa,1976).Leturepresentsavalue
func-tionforcriterionX,thenuismonotonicallyincreasingif:
[x1>x2]⇔[u
(
x1)
>u(
x2)
]. (2)https://doi.org/10.1016/j.eswa.2018.01.004
Definition2 (Keeney&Raiffa,1976).Leturepresentsavalue
func-tionforcriterionX,thenuismonotonicallydecreasingif:
[x1>x2]⇔[u
(
x1)
<u(
x2)
]. (3)Afunctionwhichisnotmonotoniciscallednon-monotonicand
mayhavedifferentshapes.Forinstance,avaluefunctionwiththe
first part increasing and the second part decreasing called
non-monotonic, by splitting of which, we have two monotonic
func-tions.
This assumption – monotonicity – however,is an
oversimpli-fication in some real-world decision-making problems. Another
simplification is the use of simple linear functions over the
en-tirerangeofacriterion.Consideringthetwoassumptions
(mono-tonicity, linearity),we usually seesimple increasing and
decreas-inglinearvalue functionsforthedecisioncriteriainMCDM
prob-lems.Theliteratureisfullofsuchapplications.Forinstance,many
of the studies reviewed in the following review papers
implic-itly adoptsuch assumptions: the MCDM applications in supplier
selection (Ho, Xu, & Dey, 2010), in infrastructure management
(Kabir, Sadiq, & Tesfamariam, 2014), in sustainable energy plan-ning(Pohekar&Ramachandran,2004),andinforestmanagement
andplanning(Ananda&Herath,2009).Whileinsomestudiesthe
use of monotonic and/or linear value function might be logical,
their use in some other applications might be unfitting. For
in-stance, Alanne, Salo, Saari, andGustafsson (2007), for evaluation
of residential energy supply systems use monotonic-linear value
functionsforall the selected evaluationcriteriaincluding “global
warmingpotential (kgCO2 m−2 a−1)”, and“acidificationpotential
(kgSO2 m−2 a−1)”.Consideringamonotonic-linearvaluefunction
forsuchcriteriaimpliesthatthedecision-makeracceptsanylevel
ofsuch harmful environmentalcriteriafor an energysupply
sys-tem. However, if the decision-maker does not accept some high
levels of such criteria (which seems logical), a piecewise linear
function might better represent the preferences of the
decision-maker(seethedecrease-levelvaluefunctioninthenextsection).
Some authors have discussed nonlinear monotonic value
functions (e.g., exponential value functions by Kirkwood, 1997;
Pratt, 1964). Others use qualitative scoring to address the
non-monotonicity(Brugha,2000;Kakeneno&Brugha,2017; O’Brien&
Brugha,2010).Wecanalsofindsomeformsofelicitingpiecewise
linear value function in Jacquet-Lagreze and Siskos (2001), and
Stewartand Janssen (2013). Some other value function
construc-tion orelicitation frameworks can be found inHerrera,
Herrera-Viedma, and Verdegay (1996), Lahdelma and Salminen (2012),
MustajokiandHämäläinen(2000),StewartandJanssen(2013),and
Yager (1988). Although in PROMETHEE we usedifferent types of
piecewisefunctionsforpairwisecomparisons(Brans,Mareschal,&
Vincke,1984),thefunctionsarenotusedto evaluatethedecision
criteria.So,despitesome efforts inliterature,thereisnoalibrary
ofsome standard piecewise linear value functions which can be
usedindifferent methodslike AHPorBWM. It isalso important
tonotethat whileinmanystudiesvaluefunctionsareelicited
ac-cordingtothepreferencedatawehavefromthedecision-maker(s),
inMCDM,usuallyweusethevaluefunctionasasubjectiveinput.
Thisimpliesthat,inMCDMmethods(exceptafewmethods,such
asUTA),thevaluefunctionisnotelicited,butanapproximationis
used.Thisalsosuggeststhattherichliteratureondeterminingand
elicitingvaluefunctionsisnotactuallyhelpingMCDMmethodsin
thisarea. In thispaper, first,a number ofpiecewise linearvalue
functionswithdifferentshapesareproposed tobeconsidered for
thedecision criteria. It is then shown, with some real-world
ex-amples,howsuchconsiderationmightchangethefinalresultsofa
decisionproblem.Acomparisonbetweensimplemonotonic linear
valuefunctions, piecewise linearvalue functions,andexponential
valuefunctionsisconducted,whichshowstheeffectivenessofthe
proposedpricewisevalue functions. Thisisasignificant
contribu-Fig. 1. Increasing value function.
tion tothis field anditis expectedto be widelyused by MCDM
applications.
Inthenextsection,somepiecewiselinearvaluefunctionsalong
withsome real-worldexamples are presented, whichis followed
bysomeremarksinSection3.InSection4,somenumerical
analy-sesareusedtoshowtheapplicabilityofconsideringtheproposed
valuefunctionsinadecisionproblem.InSection5,the
determina-tionofthevalue functionsisdiscussed.InSection 6,thepaperis
concluded,somelimitationsofthestudyarediscussed,andsome
futureresearchdirectionsareproposed.
2. Piecewiselinearvaluefunctions
Inthissection, anumberof piecewisevalue functionsare
de-fined for decisioncriteria. We providesome example casesfrom
the existing literature or practical decision-making problems to
support1 each value function.In all thefollowing valuefunctions
weconsider[dl
j, duj]asthedefineddomainforthecriterionbythe
decision-maker;xijshowstheperformanceofalternativeiwith
re-specttocriterionj;anduijshowsthevalueofalternativeiwith
re-specttocriterionj.Forinstance,ifadecision-makerwantstobuy
a carconsideringpriceasone criterion,ifall thealternatives the
decision-makerconsiders are between€17,000 and€25,000, then
thecriterionmightbedefinedforthisrange[17,000, 25,000].
2.1. Increasing
Increasing value function isperhaps the mostcommonly used
functioninMCDMapplications.Itbasicallyshowsthatasthe
crite-rionlevel,xij,increases,itsvalue,uij,increasesaswell.Itisshown
inFig.1andformulatedasfollows:
ui j=
⎧
⎨
⎩
xi j− dlj du j− dlj , dl j≤ xi j≤ d u j, 0, otherwise. (4)Forthisfunctionwecanthinkof:
• Product quality in supplier selection (Xia & Wu, 2007).
Con-sideringa set of suppliers, a buyermay always prefer a
sup-plierwithahigherproductqualitycomparedtoasupplierwith
lowerproductquality.
• Energyefficiencyinalternative-fuelbusselection(Tzeng,Lin,&
Opricovic,2005). Consideringa set ofbuses,a bus withmore
efficientfuel energymight always be preferred toa bus with
lessefficientfuelenergy.
1 It is worth-mentioning that the studies we discuss to support each value func-
tion have some theoretical or practical support for the proposed value functions. It does not, however, mean that those studies have used these value functions in their analysis.
Fig. 2. Decreasing value function.
Fig. 3. V-shape value function.
2.2. Decreasing
Decreasing value function showsthatasthe criterionlevel, xij,
increases,its value,uij,decreases.ItisshowninFig.2and
formu-latedasfollows: ui j=
⎧
⎨
⎩
du j− xi j du j− dlj , dl j≤ xi j≤ d u j, 0, otherwise. (5)Forthisfunctionwecanthinkof:
• Product pricein supplier selection (Xia & Wu, 2007).
Consid-eringasetof suppliers,a supplierwithalower product price
might always be preferred to a supplier with higher product
price.So,ahigherproductpricehasalowervalue.
• Maintenancecostinalternative-fuelbusselection(Tzengetal.,
2005).Consideringasetofbuses,abuswithlessmaintenance
costmightbepreferredtoabuswithhighermaintenancecost.
So,ahighermaintenancecostisassociatedwithalowervalue.
2.3. V-shape
V-shapevaluefunction showsthatasthecriterion level,xij,
in-creasesuptoacertainlevel,dm
j,itsvalue,uij,decreasesgradually,
andafterthatcertainlevel,dm
j,itsvalue,uij,increasesgradually.It
isshowninFig.3andformulatedasfollows:
ui j=
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
dm j − xi j dm j − d l j , dl j≤ xi j≤ dmj, xi j− dmj du j− dmj , dm j ≤ xi j≤ duj, 0, otherwise. (6)Forthisfunction,wecouldnotfindmanyexamples,anditmay
representa smallnumber ofvery particulardecision criteria. For
thisfunctionwecanthinkof:
• Relativemarketshareinselectingafirmforinvestment(Wilson
&Anell, 1999). WilsonandAnell(1999)found that for
invest-ment decision-making,firmswith low andhighmarket share
aremoredesirabletotheinvestors.Thisimpliesthatthevalue
Fig. 4. Inverted V-shape value function.
ofafirmdecreaseswhileitsmarketshareincreasesuptoa
cer-tainlevel,dm,andafterthatitsvalueincreasesagain.
• Firm size in R&D productivity(Tsai & Wang, 2005). Tsai and
Wang(2005)foundthatbothsmallandlargefirmshavehigher
R&Dproductivitycomparedtomedium-sizedfirms.Thisistrue
for both high-tech andtraditional industries. Thismeans that
therelationshipbetweensizeandvalue(measuredbyR&D
pro-ductivity)isV-shapewithaminimumlevelofvalueassignedto
acertainsizeofdm
j.
2.4.InvertedV-shape
InvertedV-shapevaluefunctionshowsthatasthecriterionlevel,
xij,increasesuptoacertain level,dmj,itsvalue,uij,increases,and
afterthatcertain level,dm
j,itsvalue, uij,decreases.Itisshownin
Fig.4andformulatedasfollows:
ui j=
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
xi j− dlj dm j − d l j , dl j≤ xi j≤ dmj, du j− xi j du j− dmj , dm j ≤ xi j≤ duj, 0, otherwise. (7)Forthisfunctionwecanthinkof:
• Commute time in selecting a job (Redmond &
Mokhtar-ian, 2001).For manypeople,the idealcommute, dm
j,is larger
than zero.Thisimpliesthatcommute timesbetweenzeroand
theoptimalcommutetime,andbetweentheoptimalcommute
timeandlargertimes,havelowervaluethantheoptimal
com-mute time for such individuals. This suggests an inverted
V-shapevaluefunction.
• Cognitive proximity in innovation partner selection
(Nooteboom, 2000). For a company there is an optimal
cognitive distanceto thepartner they areworkingon
innova-tion(dm
j).Thisimpliesthat anydistancelessthandmj orlarger
thandm
j haslessvalue.
2.5.Increase-level
Increase-levelvaluefunctionshowsthatasthecriterionlevel,xij,
increasesuptoacertainlevel,dm
j,itsvalue,uij,increases,andafter
that certain level,dm
j,its value, uij,will remain atthe maximum
level.ItisshowninFig.5andformulatedasfollows:
ui j=
⎧
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎩
xi j− dlj dm j − dlj , dl j≤ xi j≤ dmj, 1, dm j ≤ xi j≤ duj, 0, otherwise. (8)Fig. 5. Increase-level value function.
Fig. 6. Level-decrease value function.
• Fill rate in supplier selection (Chae, 2009). Although a buyer
preferssupplierswithhigherfillrate,whichimpliesthatasthe
fillrateincreasesitsvalueincreases,thebuyermightbe
indif-ferenttoanyincreaseafteracertainlevel,dm
j,asusually
buy-erspre-identifyadesirableservicelevelwhichissatisfiedbya
certainminimumlevelofsupplier’sfillrate.
• Diversity ofrestaurants inhotellocation selection (Chou,Hsu,
& Chen, 2008). In order to find the best location for an
international hotel, a decision-maker prefers locations with
more divers restaurants.However, reaching a level,dm
j, might
fullysatisfyadecision-makerimplyingthatthedecision-maker
mightnotbesensitivetoanyincreaseafterthatcertainlevel.
2.6.Level-decrease
Level-decrease value function showsthat asthe criterion level,
xij,increasesuptoacertainlevel,dm
j,itsvalue,uij,remainsat
max-imumlevel,andafterthatcertainlevel,dm
j,itsvalue,uij,decreases
gradually.ItisshowninFig.6andformulatedasfollows:
ui j=
⎧
⎪
⎪
⎨
⎪
⎪
⎩
1, dl j≤ xi j≤ dmj, du j− xi j du j− dmj , dm j ≤ xi j≤ duj, 0, otherwise. (9)Forthisfunctionwecanthinkof:
• Distanceinselectingauniversity(Carricoetal.,1997).Whilea
student prefersa closeruniversity toa farther university, this
preference might start after a certain distance, dm
j, implying
thatanydistancebetween[dl
j,dmj]isoptimalandindifferentfor
thestudent.
• Leadtime insupplierselection(Çebi&Otay,2016).Althougha
supplierwithashorterleadtime ispreferred,iftheleadtime
is in a limit such that it doesnot negatively affect the
com-pany’s production, the company might then be indifferent to
thatrange.
Fig. 7. Level-increase value function.
Fig. 8. Decrease-level value function.
2.7. Level-increase
Level-increase value function shows that asthe criterion level,
xij,increasesuptoacertainlevel,dmj,itsvalue,uij,remainsat
min-imumlevel,andafterthatcertainlevel,dm
j,itsvalue,uij,increases.
ItisshowninFig.7andformulatedasfollows:
ui j=
⎧
⎪
⎪
⎨
⎪
⎪
⎩
0, dl j≤ xi j≤ dmj, xi j− dmj du j− dmj , dm j ≤ xi j≤ duj, 0, otherwise. (10)Forthisfunctionwecanthinkof:
• Leveloftrustinmakingabuyer–supplierrelationship(Ploetner
&Ehret,2006).Trustincreasesthelevelofpartnershipbetween
abuyerandasupplier,howeveritisonlyeffectiveaftera
cer-tainthreshold,dm
j.
• Level of relational satisfaction in evaluating quality
commu-nication in marriage (Montgomery, 1981). Below a minimum
levelofrelationalsatisfaction,dm
j,qualitycommunication
can-nottakeplacethusresultsinminimumvalue.
2.8. Decrease-level
Decrease-level value function showsthat asthe criterion level,
xij,increasesuptoacertainlevel,dmj,its value,uij,decreases,and
afterthatcertain leveldm
j,itsvalue,uij,willremain atminimum.
ItisshowninFig.8andformulatedasfollows:
ui j=
⎧
⎪
⎪
⎨
⎪
⎪
⎩
dm j − xi j dm j − dlj , dl j≤ xi j≤ dmj, 0, dm j ≤ xi j≤ duj, 0, otherwise. (11)Forthisfunctionwecanthinkof:
• Carbon emission intransportation mode selection (Hoen, Tan,
Fransoo, & van Houtum, 2014). In selecting a transportation
mode,themorethecarbonemissionbythemode,thelessthe
Fig. 9. Increasing stepwise value function.
zerovaluetoa modewithcarbonemissionhigherthana
cer-tainlevel,dm
j.
• Distancewhen selectinga school(Frenette,2004). Ithasbeen
shownthat the longerthe distanceto the school thelessthe
preferencetoattendthatschool.It isalsoclearthat,forsome
people,thereisnovalueafteracertaindistance,dm
j.
2.9. Increasingstepwise
Increasing stepwise value function shows that asthe criterion
level, xij, increasesup to a certain level,dmj,its value remains at
a certain level, u0, andafter that certain level, dmj, its value, uij,
jumpstoahigherlevel(maximum)andremainsatthemaximum.
ItisshowninFig.9andformulatedasfollows:
ui j=
⎧
⎨
⎩
u0, dlj≤ xi j≤ dmj, 1, dm j ≤ xi j≤ duj, 0, otherwise. (12) where0<u0<1.Forthisfunctionwecanthinkof:
• Suppliers capabilities in supplier segmentation (Rezaei &
Ortt, 2012). Suppliers of a company are evaluated based on
theircapabilities,andthensegmentedbasedontwolevels(low
andhigh)withrespecttotheir capabilities.Assuch asupplier
scored betweendl
j anddmj isconsidered asa low-level
capa-bilitiessupplier whileasupplier scoredbetweendm
j andduj is
consideredasahigh-levelcapabilitiessupplier.
• Symmetry in selecting a close type of partnership
(Lambert, Emmelhainz, & Gardner, 1996). In order to have
asuccessful relationshipbetweensupplychainpartners, there
should be some demographical similarities (for instance, in
terms of brand image, productivity) between them. So, more
symmetry means closer relationship. However, if we consider
two levels of closeness, it is clear that for some level of
symmetrythevalueofclosenessremainsthesame.
For the increasing stepwise value function, a criterion might
havemorethanonejump.Forinstance,ifadecision-makerwants
toconsider threelevelslow,medium,andhighwhensegmenting
thesupplierswithrespecttotheir capabilities,thenan increasing
stepwisefunctionwithtwojumpsshouldbedefinedforthis
crite-rion.Thefollowingvaluefunctionisageneralincreasing stepwise
valuefunctionwithkjumps.
ui j=
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
u0, dlj≤ xi j≤ dmj1, u1, dmj1≤ xi j≤ dmj2, . . . 1, dmk j ≤ xi j≤ d u j, 0, otherwise. (13) where0<u0<u1<…<1.Fig. 10. Decreasing stepwise value function.
2.10.Decreasingstepwise
Decreasing stepwise value function shows that asthe criterion
level,xij,increasesuptoa certainlevel,dmj,its value,uij,remains
atathemaximumlevel,andafterthatcertainlevel,dm
j,itsvalue,
uij,jumpsdowntoalowerlevel,u0,andremainsatthatlevel.Itis
showninFig.10andformulatedasfollows:
ui j=
⎧
⎨
⎩
1, dl j≤ xi j≤ dmj, u0, dmj ≤ xi j≤ duj, 0, otherwise. (14) where0<u0 <1.Forthisfunctionwecanthinkof:
• Considering supply risk in portfolio modeling (Kraljic, 1983).
For a company, an item witha higherlevel of risk results in
lessvalue,however,duetoportfoliomodeling,thereisno
dif-ferencebetweenalllevelsofriskinthedomain[dl
j, dmj].
Simi-larly,alllevelsofriskinthedomain[dm
j, duj]resultinthesame
value.
• Delay in logistics service provider selection (Qi, 2015). Some
companies consider stepwise value function for delay in
de-liveringtheitemsbya logisticsserviceprovider,which means
thatthevalueofthatproviderdecreaseswhendelayincreases,
howeveritisconstantwithincertainintervals.
For the decreasing stepwise function, a criterion might have
morethanonejump.Thefollowingvaluefunctionisageneral
de-creasingstepwisefunctionwithkjumps.
ui j=
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
1, dl j≤ xi j≤ dmj1, u1, dmj1≤ xi j≤ dmj2, . . . uk, dmkj ≤ xi j≤ d u j, 0, otherwise. (15) where0<uk<…<u1 <1.3. Someremarksonthevaluefunctions
Here, a number of remarks are discussed, shedding light on
some aspects of the proposed value functions, which might be
usedinreal-worldapplications.
Remark1. Shape andparameters ofa valuefunction is
decision-maker-dependent,implyingthat(i)whileadecision-maker
consid-ers, for instance, a level-increasing function for the size of
gar-den when buying a house, another decision-makerconsiders an
increasing stepwise function, and (ii) while two decision-makers
considerincreasingstepwisefunctionforthesizeofgardenwhen
buyinga house, the parameters they consider fortheir functions
(
dlFig. 11. Increasing-level-decreasing value function.
Table 1
Suppliers performance with respect to different decision criteria ( x ij ).
Criteria
Supplier Quality Price ( €/item) Trust CO 2 (g/item) Delivery (day)
1 85 27 4 10 0 0 3 2 90 28 2 1500 4 3 80 26 5 20 0 0 3 4 75 25 5 10 0 0 2 5 95 29 7 1700 3 6 99 30 6 20 0 0 1
Remark2. Adecision-makermightconsidera hybridvalue
func-tion fora criterion. For instance, a criterion might be
character-ized withan increasing-level-decreasing, which is a combination
ofincreasing-levelandlevel-decreasing.Thisfunctioncan alsobe
considered as a special form of inverted V-shape function. It is
showninFig.11andformulatedasfollows:
ui j=
⎧
⎪
⎪
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎪
⎩
xi j− dlj dm j − dlj , dl j≤ xi j≤ dmj1, 1, dm1≤ x i j≤ dm2, du j− xi j du j− dmj , dm2≤ x i j≤ duj, 0, otherwise. (16)Forinstance,adecision-makerhastoselectthebestR&D
part-neramong10partners.Oneofthecriteriaisdistanceandthe
com-panygivesless preferenceto very closeorvery distant partners,
which are distributed in the range [10km, 2000km]. The
com-panyconsidersanoptimaldistanceof[200km, 500km].This
im-pliesthatdistancefollowsan‘increasing-level-decreasing’function
forthisdecision-maker:[dl
j, d m1 j ,d m2 j ,d u j]=[10, 200, 500, 2000].
4. Numericalandcomparisonanalyses
In this section, we show how to incorporate the proposed
piecewise value functionsinto account when applyingan MCDM
method,andweshowthattheresultsmightbedifferentwhenwe
considertheproposedpiecewisevaluefunctions.
WeconsideranMCDMproblem,whereabuyershouldselecta
supplierfromamongsixqualifiedsuppliersconsideringfive
crite-ria:qualitywhichismeasuredby1−
α
,whereα
showsthelot-sizeaverageimperfectrate;price(euro)peritem;trust,whichis
mea-suredbya Likertscale (1:verylow to 7:very high);CO2 (gram)
peritem;delivery(day),theamountoftimewhichtakestodeliver
itemsfromthesuppliertothebuyer(allthecriteriaarecontinuous
excepttrust). Table1showsthe performanceof thesixsuppliers
withrespecttothefivecriteria.
Thebuyerhasusedanelicitation method2 tofindtheweights
whichareasfollows:
w∗quality =0.20,w∗price=0.30, w∗trust=0.27, w∗CO2 =0.08, w
∗
delivery=0.15.
And we assume that the decision-maker considers piecewise
valuefunctionsforthesecriteria(see,Table2).
So, as can be seen from Table 2, the decision-maker
consid-ersalevel-increaselinearfunctionforqualitywiththelowestand
highestvalues of0 and100, respectively.For thedecision-maker
anynumberbelow85hasno valueatall. Forcriterionprice, the
decision-makergivesthehighestvalue toanypricebelow15
(al-though in the existing set of suppliersthere is no supplier with
a price within thisrange), after which thevalue decreases till it
reachesto themaximum priceof30.Forcriterion trust whichis
measured using a Likert scale (1:very low to 7: very high),any
number less than 3 has no value for the decision-maker, while
thevalue gradually increasesbetween3 and7.ForCO2 emission,
thereisadecreasingvaluefunctionfrom0to1500gperitem,
af-terwhichtill2000g,allthenumbershavezerovalue.Finally,for
deliverythere isa simpledecreasingfunctionwithminimumand
maximumvaluesof0and5days.
Byusingthefollowingequation, wecan findtheoverall value
ofeachsupplierandthenrankthemtofindthebestsupplier.
Ui= n
j=1
wjui j (17)
where,uij is the value of the performance of supplier i with
re-specttocriterionj (usingtheequationsinTable2forthedatain
Table1),andwjistheweightofcriterionjasfollows:
w∗quality=0.20,w∗price=0.30, w∗trust=0.27,
w∗CO2=0.08,w∗delivery=0.15.
The value scores and the aggregated values are presented in
Table3(seealsoFig.13forthefinalresults).
AscanbeseenfromTable3,supplier6withthegreatestoverall
valueof0.51isrankedasthefirstsupplier.Suppliers5,4,3,1,and
2arerankedinthenextplaces.
4.1.Comparingwiththesimplelinearvaluefunctions
Inexistingliterature,consideringthenatureofthecriteria,the
valuesarecalculated,forinstance,usingthe followingsimple
lin-earvaluefunction:
ui j=
⎧
⎪
⎪
⎨
⎪
⎪
⎩
xi j− dlj du j− d l j, ifmorexi j ismoredesirable
(
suchasquality)
,du j− xi j du j− d l j
, ifmorexi jislessdesirable
(
suchasprice)
.(18)
Eq.(18)is usedtofindthevaluesofthecriteriaforeach
sup-plier usingthe data inTable 1. Considering the criteria weights,
anduij (Eq.(18)forthe datainTable 1) usingEq.(17)the value
scoresandalsotheaggregatedoverallscoreofeachalternativecan
becalculatedwhichareshowninTable4(seealsoFig.13forthe
finalresults).
InTable 4 itis assumedthat quality andtrust arecriteriafor
whichthehigherthebetter,whilefortheothercriteria(price,CO2,
anddelivery),thelowerthebetter.Infact,weconsidersimple
lin-ear functions(increasinganddecreasingrespectively)forthe two
groupsofcriteria.
2 Please note that we report some weights for the criteria as the aim of the study
Table 2
Piecewise value functions.
Shape Value function
u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 0 , d l j ≤ x i j ≤ d mj , x i j − d mj d u j − d mj d m j ≤ x i j ≤ d uj , 0 , otherwise . = ⎧ ⎪ ⎨ ⎪ ⎩ 0 , 0 ≤ x i j ≤ 85 , x i j − 85 100 − 85 , 85 ≤ x i j ≤ 100 , 0 , otherwise . u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 1 , d l j ≤ x ≤ d mj , d u j − x i j d u j − d mj , d m j ≤ x ≤ d uj , 0 , otherwise . = ⎧ ⎪ ⎨ ⎪ ⎩ 1 , 13 ≤ x i j ≤ 15 , 30 − x i j 30 − 15 , 15 ≤ x i j ≤ 30 , 0 , otherwise . u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 0 , d l j ≤ x i j ≤ d mj , x i j − d mj d u j − d mj , d m j ≤ x i j ≤ d uj , 0 , otherwise . = ⎧ ⎪ ⎨ ⎪ ⎩ 0 , 1 ≤ x i j ≤ 3 , x i j − 3 7 − 3 , 3 ≤ x i j ≤ 7 , 0 , otherwise . u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ d m j − x i j d m j − d lj , d l j ≤ x i j ≤ d mj , 0 , d m j ≤ x i j ≤ d uj , 0 , otherwise . = ⎧ ⎪ ⎨ ⎪ ⎩ 1500 − x i j 1500 − 0 , 0 ≤ x i j ≤ 1500 , 0 , 1500 ≤ x i j ≤ 2000 , 0 , otherwise . u i j = ⎧ ⎨ ⎩ d u j − x i j d u j − d lj , d l j ≤ x i j ≤ d uj , 0 , otherwise . = 5 − x i j 5 − 0 , 0 ≤ x i j ≤ 5 , 0 , otherwise .
According toTable 4,the best supplieris supplier 4,whichis
ranked as the3rd one considering the piecewise value functions
(Table3). The rankingof theother suppliersis alsodifferent.So,
ascanbeseen,suchdifferencesareassociatedtothewaywe
cal-culatethevalueofthecriteria.Ifwelookatthecriteriontrust,for
instance(Table3),wecanseethatonlythenumbersgreaterthan
3canbeusedforcompensatingtheothercriteria.Inotherwords,
thevalues1,2and3forthiscriterionhavenoselectionpower.No
suppliercancompensateitsweaknessinothercriteriabyhavinga
valuebetween1and3fortrust.However,suchimportantissueis
entirelyignoredinthesimplewayofdeterminingthevalue
Table 3
Value scores, u ij , and the aggregated overall score considering the proposed piecewise value
functions.
Supplier Quality Price Trust CO 2 Delivery Aggregated value Rank
1 0.00 0.20 0.25 0.50 0.40 0.23 5 2 0.33 0.13 0.00 0.00 0.20 0.14 6 3 0.00 0.27 0.50 0.00 0.40 0.28 4 4 0.00 0.33 0.50 0.50 0.60 0.37 3 5 0.67 0.07 1.00 0.00 0.40 0.48 2 6 0.93 0.00 0.75 0.00 0.80 0.51 1 Table 4
Value scores, u ij , and the aggregated overall scores considering the simple linear value func-
tions.
Supplier Quality Price Trust CO 2 Delivery Aggregated value Rank
1 0.42 0.60 0.40 1.00 0.33 0.50 4 2 0.63 0.40 0.00 0.50 0.00 0.29 6 3 0.21 0.80 0.60 0.00 0.33 0.49 5 4 0.00 1.00 0.60 1.00 0.67 0.64 1 5 0.83 0.20 1.00 0.30 0.33 0.57 2 6 1.00 0.00 0.80 0.00 1.00 0.57 3
Fig. 12. Exponential value functions.
isevenofahigherimportanceforcompensatorymethodssuchas
AHPandBWM.
4.2.Comparingwiththeexponentialvaluefunctions
Another importantwayto approximate thevalue functions in
practiceistheuseofexponentialvaluefunctions(Kirkwood, 1997;
Pratt, 1964). The exponential value functions can specifically be
used when the preferences are monotonically increasing or
de-creasing. Although this approach is not popular in MCDM
do-main,andwewere notabletofindanyapplicationofthesevalue
functionsparticularly in MCDM field, we would like to compare
ourresults to the results ofapplying thesefunctions, whichare,
to some extent, close to some of our proposed piecewise value
functions(suchaslevel-increase,level-decrease,increase-level,and
decrease-level).Usingthe samenotationsasbeforeand
consider-ingashape parameter
ρ
whichiscalled‘risktolerance’,amono-tonically increasing exponential value function can be shown as
follows: ui j=
⎧
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎩
1− exp−xi j− dlj /
ρ
1− exp−du j− dlj /
ρ
,ρ
=Infinity xi j− dlj du j− dlj , otherwise. (19)Amonotonically decreasing exponential value function can be
shownasfollows: ui j=
⎧
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎩
1− exp−du j− xi j /
ρ
1− exp−du j− dlj /
ρ
,ρ
=Infinity du j− xi j du j− dlj , otherwise. (20)Fig. 12 showsthe monotonically increasing exponential value
functions(fordifferentvaluesof
ρ
)(a),andthemonotonicallyde-creasingexponentialvaluefunctions(fordifferentvaluesof
ρ
)(b).Risk-averse decision-makershave
ρ
> 0 (hill-like functionsinFig.12),whilerisk-seekingdecision-makershave
ρ
<0(bowl-likefunctionsinFig.12).
ρ
=Infinity(straight-lineinFig.12)showsthevaluefortheriskneutraldecision-makers.Infact,
ρ
=Infinitypro-ducesthesimplelinearvaluefunctionswhichareverypopularin
MCDMfield.
In order to do the comparison analysis, we use exponential
value functions for the criteria of the aforementioned example
(Table 1) to check the similarities and differences. To make a
faircomparison,we tryto generate3 the corresponding
exponen-tial value functionsof thepiecewise value functions(Table2) as
3 To see how these value functions are elicited considering the decision-maker
closeaspossible.Forquality,amonotonicallyincreasing
exponen-tialvaluefunctionwithnegative
ρ
wouldbeappropriate.Forpriceamonotonicallydecreasingexponentialvaluefunctionwitha
posi-tive
ρ
,fortrust,amonotonicallyincreasingexponentialvaluefunc-tion with a negative
ρ
, for CO2, a monotonically decreasingex-ponential value function with a negative
ρ
, and, finally, forde-liverya monotonicallydecreasingexponentialvaluefunction with
ρ
=Infinitywouldbesuitable.Table5showsthefunctions,wherefunctionswithdifferent
ρ
sareshownandamoresuitable oneisshowninbold.
Using theexponential value functionsof Table5, forthe data
ofTable 1,we get thevalue scores andthe aggregatedvaluesas
presentedinTable6(seealsoFig.13forthefinalresults).
Table 5
Exponential value functions.
Shape Function u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 1 − exp [ −( x i j − d lj) /ρ] 1 − exp[ −( d u j − d lj) /ρ] , ρ = Infinity x i j − d lj d u j − d lj , otherwise . = ⎧ ⎨ ⎩ 1 − exp [ −( x i j − 0 ) /ρ] 1 − exp[ −( 100 − 0 ) /ρ] , ρ = Infinity x i j − 0 100 − 0 , otherwise . u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 1 − exp [ −( d u j − x i j) /ρ] 1 − exp[ −( d u j − d lj) /ρ] , ρ = Infinity d u j − x i j d u j − d lj , otherwise . = ⎧ ⎨ ⎩ 1 − exp [ −( 30 − x i j) /ρ] 1 − exp[ −( 30 − 0 ) /ρ] , ρ = Infinity 30 − x i j 30 − 0 , otherwise . u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 1 − exp [ −( x i j − d lj) /ρ] 1 − exp[ −( d u j − d lj) /ρ] , ρ = Infinity x i j − d lj d u j − d lj , otherwise . = ⎧ ⎨ ⎩ 1 − exp [ −( x i j − 1 ) /ρ] 1 − exp[ −( 7 − 1 ) /ρ] , ρ = Infinity x i j − 1 7 − 1 , otherwise . u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 1 − exp [ −( d u j − x i j) /ρ] 1 − exp[ −( d u j − d lj) /ρ] , ρ = Infinity d u j − x i j d u j − d lj , otherwise . = ⎧ ⎨ ⎩ 1 − exp [ −( 20 0 0 − x i j) /ρ] 1 − exp[ −( 20 0 0 − 1500 ) /ρ] , ρ = Infinity 20 0 0 − x i j 20 0 0 − 0 , otherwise .
Table 5 ( continued ) Shape Function u i j = ⎧ ⎪ ⎪ ⎨ ⎪ ⎪ ⎩ 1 − exp [ −( d u j − x i j) /ρ] 1 − exp[ −( d u j − d lj) /ρ] , Infinity d u j − x i j d u j − d lj , otherwise . = ⎧ ⎨ ⎩ 1 − exp [ −( 5 − x i j) /ρ] 1 − exp[ −( 5 − 0 ) /ρ] , ρ = Infinity 5 − x i j 5 − 0 , otherwise . Table 6
Value scores, u ij , and the aggregated overall scores considering the exponential value func-
tions.
Supplier Quality Price Trust CO 2 Delivery Aggregated value Rank
1 0.85 0.45 0.05 0.08 0.40 0.22 5 2 0.90 0.33 0.00 0.02 0.20 0.16 6 3 0.80 0.55 0.13 0.00 0.40 0.26 4 4 0.75 0.63 0.13 0.08 0.60 0.32 3 5 0.95 0.18 1.00 0.00 0.40 0.46 1 6 0.99 0.00 0.37 0.00 0.80 0.38 2
Fig. 13. Final results using three types of value functions.
AscanbeseenfromTable6,supplier5withthegreatest
over-allvalue of0.46isrankedasthefirst supplier,which isdifferent
fromwhat we get from the proposed piecewise value functions
(Table3).Whilesupplier5wasrankedthe2ndbasedonour
pro-posedvaluefunctions, usingthe exponentialvalue functions, this
supplierbecomesnumber2. Other suppliers(1,2,3,4) have the
samerankingbasedonthetwoapproaches.Thedifferencesare
ob-viouslyassociatedtothewaywegetthevaluesofthecriteria.We
alsocheckedsomeother close
ρ
valuesforthe exponentialvaluefunctions. There are some changes in the aggregatedvalues, yet,
therankingisthesame.
Asitcanbeseen,theresultsofthetwoapproaches(piecewise
valuefunctionsandexponentialfunctionsaremuchclosertoeach
other than tothe resultsof theregular simple linearvalue
func-tions).
Our observation is that the exponential value functions can
play a role close to a number of proposed value functions in
thisstudysuchasincrease-level,decrease-level,level-increase,and
level-decrease. In order to make an exponential value functions
closetooneofthementionedproposedvaluefunctionsweshould
choose
ρ
valuesclose to zero.If we try to make theρ
ascloseFig. 14. Fitting an exponential value function to a level-decrease value function.
otherpartbecomesverysteepandnotrepresentative.Ontheother
hand,ifwewanttochoosea
ρ
value whichbetterrepresentstheslopeofthefunctionforthe“increase” orthe“decrease” part,then
itisimpossibletocoverthe“level” partofthevaluefunction
prop-erly.
For example, let usconsider the criterion price again. In Fig.
14,itcanbeseenthat,ifweconsider
ρ
=1,thelevelpartisfullycovered, but the decrease part ofthe exponential value function
is verymuch differentfromthe decreasepartofthe linear
func-tion.Evenifwechoose
ρ
=3, whichdoesnotfullycoverthelevelpartoftheproposedfunction,thedecreasepartisreallydifferent.
On the other hand,we can find
ρ
=8 as a close one to thede-creasepartof thelinearfunction, butthistime itis notpossible
tocoverthelevelpartproperly.So,althoughaverygood
approx-imation,theexponentialvaluefunctionsmightnotbesuitable for
casesinwhichadecision-makerhasaclearvalue-indifference
in-terval (alevel part) for a criterion. However, thesefunctions are
indeedsuitablewhenthedecision-makerhasdifferentpreferences
onthelowerandontheupperpartsofthecriterionmeasure.
Fromthefigurewecanalsoseethatwhilewecouldmakethe
two piecewise and exponential value functions, to some degree,
closeto each other,they are toodifferentfromthe simplelinear
valuefunction.Itisalsoclearthatnoneofthesimplelinearvalue
functionsortheexponential valuefunctionscanrepresentthe
V-orinvertedV-shapevaluefunctions.
Asa generalconclusion,we thinkthattheproposed piecewise
linear functions havetwo salient features: (i) simplicity; and(ii)
representativeness.Thatis,itiseasytoworkwithlinearfunctions,
anditiseasyforapractitionertofindamorerepresentative
func-tionfromtheproposedlibraryofthepricewisevaluefunctionsfor
a particularcriterion.The cut-off pointscanalsobe estimatedby
the decision-maker.The simple monotonic-linear value functions,
which are dominant inexisting literature, are very simple.
How-ever, theymight notbe representative insome cases.Finally,the
exponentialvaluefunctionsmighthaveabetterrepresentativeness
(comparedtothe simplemonotonic-linearvalue functions).
How-ever,theyarenotsimple.Workingwithnon-linerfunctionsisnot
easyforpractitioner,and,moreimportantly,itisverydifficultfora
practitionertoestimateavalue for
ρ
(the shapeparameteroftheexponential valuefunctions),asitcannotbe easilyinterpreted by
apractitioner(pleasenotethatweconsideravaluefunctionasan
inputforanMCDMprobleminthisstudy).
5. Determiningthevaluefunctions
One ofthe big challenges inreal-world decision-making is to
find a proper value function for a decision criterion. This,
per-haps, has been one of the main reasons why the use of simple
linear value functions in multi-criteria decision-making is
domi-nant. The linear value functions are easy for modeling purposes
andcan, tosome extent,representthe reality.More complicated
valuefunctions,althoughmightbeclosertorealityofthe
decision-maker’spreferences, are moredifficultto be elicitedandare
dif-ficult for modeling purposes. We refer the interested readers to
someexistingproceduresforidentifyingvaluefunctions(Fishburn,
1967;Keeney & Nair, 1976; Kirkwood, 1997;Pratt, 1964; Stewart & Janssen, 2013). We think that the proposed value functions in
thispaperdonothavethedisadvantageofnonlinearityandatthe
same time have the advantagesof being closer to the real
pref-erencesofthedecision-makerasthey providesomediversityand
flexibility in modeling the functions. As we do not consider the
nonlinearityofthe value functionswedo not usetheconcept of
risk tolerance in determining the value functions as it has been
used by others.We rather propose a simple procedure, which is
morepractical.
A decision analyst, could first show the value functions in
Table 2 to the decision-maker to see which one most suits the
preference structure of the decision-maker. Once the
decision-makerselectsa particularvalue function, theother details ofthe
function, such as the lower bound, the upper bound and the
thresholds canbe determined. We should highlightagain that in
mostMCDMmethods,thevaluefunctionisnotelicited.Itisrather
simply assumed to have a particular shape, and this is why we
thinkhavingapre-specifiedsetofstandardvaluefunctionswhich
canbeusedassubjectiveapproximationoftherealpreferencesof
thedecision-makercanmakeasignificantimpactontheresults.
6. Conclusion,limitationsandfutureresearch
This study proposes a set of piecewise value functions for
multi-criteriadecision-making(MCDM)problems.Whilethe
exist-ingapplicationsofMCDMmethodsusuallyusetwogeneralsimple
increasing and decreasing linear value functions, this study
pro-vides severalreal-world examples to support the applicability of
someotherformsofvaluefunctionsforthecriteriausedinMCDM.
Itis alsoexplicated how, insome decisionproblems,a
combina-tion of two or more value functions can be used for a
dif-ferentMCDM methods indifferent decisionproblems. A
numeri-calexampleofsupplierselectionproblem(includingacomparison
betweensimple monotonic linear value functions, piecewise
lin-earvaluefunctions,andexponentialvaluefunctions)showedhow
theuseofthe proposed valuefunctionscould affectthe final
re-sults.Consideringthesevaluefunctionscouldbetterrepresentthe
realpreferencesofthedecision-maker.Itcanalsohelpreducethe
inappropriatecompensations ofthedecisioncriteria, forinstance,
throughusingalevel-increasingfunctionwhichassignszerovalue
toany value ofthe criterion below a certain threshold.The
pro-posedvalue functionsare presented in a generalform such that
they can be tailor-made for a specific decision-maker. Thatis to
say,notonlyitispossiblefortwodifferentdecision-makerstouse
twodifferentvalue functionsfora singlecriterion.It isalso
pos-sibleto use different domain (e.g.min and max) values forthat
particularvaluefunction.
Despite the advantages of the proposed value functions, they
havesomelimitations.Althoughtheproposedvaluefunctions
con-sidersomereal-worldfeaturesofthedecisioncriteria,theyare
lin-ear which might be, to some degree, a simplification. We think
thatthe biggerproblemin existingliterature of MCDMis
mono-tonicity assumption and not linearity assumption. Nevertheless,
moreresearchneedstobeconductedtoempiricallyfindtheshare
ofeach.Furthermore,toformulatethedecisioncriteriaoneshould
payenoughattentiontochecktherealcontributionofthedecision
criterion into the ultimate goal of the decision-making problem.
Forinstance,if acriterion contributesto anothercriterion which
hasarealroleinmakingthedecision,oneshouldexcludethe
ini-tialone. Foradetaileddiscussiononthismatter, interested
read-ersarereferred toBrugha(1998).Oneinterestingfuturedirection
wouldbetoapplytheproposedvaluefunctionsinsomereal-world
MCDMproblemsandcomparetheirfitnesstotheothervalue
func-tions.Inthisregard,findingamoresystematicapproachto
deter-minethe valuefunctionsin practicewouldbe alsovery
interest-ing.Itwouldbealsointerestingtostudythecasesinwhichthere
are more than one decision-maker. As different decision-makers
maychoose differentvalue functions, different domains, and
dif-ferent thresholds for a single criterion, proposing a way to find
the final output of the MCDM problem for the group would be
an interesting future research. Finally, finding a sensitivity
anal-ysisforthe proposed value functions isrecommended.
Consider-ingthestudiesofBertschandFichtner(2016),Bertsch,Treitz,
Gel-dermann, and Rentz (2007), Insua and French (1991), Wulf and Bertsch(2017)couldgiveinterestingideastomakesuchsensitivity
analysisframework.
References
Alanne, K. , Salo, A. , Saari, A. , & Gustafsson, S.-I. (2007). Multi-criteria evaluation of residential energy supply systems. Energy and Buildings, 39 , 1218–1226 . Ananda, J. , & Herath, G. (2009). A critical review of multi-criteria decision making
methods with special reference to forest management and planning. Ecological Economics, 68 , 2535–2548 .
Bertsch, V. , & Fichtner, W. (2016). A participatory multi-criteria approach for power generation and transmission planning. Annals of Operations Research, 245 , 177–207 .
Bertsch, V. , Treitz, M. , Geldermann, J. , & Rentz, O. (2007). Sensitivity analyses in multi-attribute decision support for off-site nuclear emergency and recovery management. International Journal of Energy Sector Management, 1 , 342–365 . Brans, J. P. , Mareschal, B. , Vincke, P. , & Brans, J. P. (1984). PROMETHEE: A new fam-
ily of outranking methods in multicriteria analysis. In Proceedings of the 1984 conference of the international federation of operational research societies (IFORS) (pp. 477–490). Amsterdam: North Holland. 84 .
Brugha, C. M. (1998). Structuring and weighting criteria in multi criteria deci- sion making (MCDM). In Trends in multicriteria decision making (pp. 229–242). Springer .
Brugha, C. M. (20 0 0). Relative measurement and the power function. European Jour- nal of Operational Research, 121 , 627–640 .
Carrico, C. S. , Hogan, S. M. , Dyson, R. G. , & Athanassopoulos, A. D. (1997). Data envel- opment analysis and university selection. The Journal of the Operational Research Society, 48 , 1163–1177 .
Çebi, F. , & Otay, ˙I. (2016). A two-stage fuzzy approach for supplier evaluation and order allocation problem with quantity discounts and lead time. Information Sci- ences, 339 , 143–157 .
Chae, B. (2009). Developing key performance indicators for supply chain: An indus- try perspective. Supply Chain Management: An International Journal, 14 , 422–428 . Chou, T.-Y. , Hsu, C.-L. , & Chen, M.-C. (2008). A fuzzy multi-criteria decision model for international tourist hotels location selection. International Journal of Hospi- tality Management, 27 , 293–301 .
Edwards, W. (1977). How to use multiattribute utility measurement for social deci- sion making. IEEE Transactions on Systems, Man and Cybernetics, 7 , 326–340 . Fishburn, P. C. (1967). Methods of estimating additive utilities. Management Science,
13 , 435–453 .
Frenette, M. (2004). Access to college and university: Does distance to school mat- ter. Canadian Public Policy/Analyse de Politiques, 30 , 427–443 .
Herrera, F. , Herrera-Viedma, E. , & Verdegay, J. (1996). Direct approach processes in group decision making using linguistic OWA operators. Fuzzy Sets and Systems, 79 , 175–190 .
Ho, W. , Xu, X. , & Dey, P. K. (2010). Multi-criteria decision making approaches for supplier evaluation and selection: A literature review. European Journal of Oper- ational Research, 202 , 16–24 .
Hoen, K. , Tan, T. , Fransoo, J. , & van Houtum, G. (2014). Effect of carbon emission reg- ulations on transport mode selection under stochastic demand. Flexible Services and Manufacturing Journal, 26 , 170–195 .
Insua, D. R. , & French, S. (1991). A framework for sensitivity analysis in discrete multi-objective decision-making. European Journal of Operational Research, 54 , 176–190 .
Jacquet-Lagreze, E. , & Siskos, Y. (2001). Preference disaggregation: 20 years of MCDA experience. European Journal of Operational Research, 130 , 233–245 .
Kabir, G. , Sadiq, R. , & Tesfamariam, S. (2014). A review of multi-criteria decision– making methods for infrastructure management. Structure and Infrastructure En- gineering, 10 , 1176–1210 .
Kakeneno, J. R. , & Brugha, C. M. (2017). Usability of nomology-based methodologies in supporting problem structuring across cultures: The case of participatory de- cision-making in Tanzania rural communities. Central European Journal of Oper- ations Research, 25 , 393–415 .
Keeney, R. L., & Nair, K. (1976). Evaluating potential nuclear power plant sites in the Pacific Northwest using decision analysis, IIASA Professional Paper. IIASA, Laxenburg, Austria: PP-76-001.
Keeney, R. L. , & Raiffa, H. (1976). Decisions with multiple objectives: preferences and value tradeoffs . USA: John Wiley & Sons, Inc .
Kirkwood, C. W. (1997). Strategic decision making . California, USA: Wadsworth Pub- lishing Company .
Kraljic, P. (1983). Purchasing must become supply management. Harvard Business Review, 61 , 109–117 .
Lahdelma, R. , & Salminen, P. (2012). The shape of the utility or value function in stochastic multicriteria acceptability analysis. OR Spectrum, 34 , 785–802 . Lambert, D. M. , Emmelhainz, M. A. , & Gardner, J. T. (1996). Developing and imple-
menting supply chain partnerships. The International Journal of Logistics Manage- ment, 7 , 1–18 .
Montgomery, B. M. (1981). The form and function of quality communication in mar- riage. Family Relations, 30 , 21–30 .
Mustajoki, J. , & Hämäläinen, R. P. (20 0 0). Web-HIPRE: Global decision support by value tree and AHP analysis. INFOR: Information Systems and Operational Re- search, 38 , 208–220 .
Nooteboom, B. (20 0 0). Learning and innovation in organizations and economies . Ox- ford: University Press .
O’Brien, D. B. , & Brugha, C. M. (2010). Adapting and refining in multi-criteria deci- sion-making. Journal of the Operational Research Society, 61 , 756–767 . Ploetner, O. , & Ehret, M. (2006). From relationships to partnerships—New forms
of cooperation between buyer and seller. Industrial Marketing Management, 35 , 4–9 .
Pohekar, S. , & Ramachandran, M. (2004). Application of multi-criteria decision mak- ing to sustainable energy planning—a review. Renewable and Sustainable Energy Reviews, 8 , 365–381 .
Pratt, J. W. (1964). Risk aversion in the small and in the large. Econometrica, 32 , 122–136 .
Qi, X. (2015). Disruption management for liner shipping. In Handbook of ocean con- tainer transport logistics (pp. 231–249). Springer .
Redmond, L. S. , & Mokhtarian, P. L. (2001). The positive utility of the commute: Modeling ideal commute time and relative desired commute amount. Trans- portation, 28 , 179–205 .
Rezaei, J. (2015). Best-worst multi-criteria decision-making method. Omega, 53 , 49–57 .
Rezaei, J. (2016). Best-worst multi-criteria decision-making method: Some proper- ties and a linear model. Omega, 64 , 126–130 .
Rezaei, J. , & Ortt, R. (2012). A multi-variable approach to supplier segmentation. In- ternational Journal of Production Research, 50 , 4593–4611 .
Roy, B. (1968). Classement et choix en présence de points de vue multiples (la méthode ELECTRE). RIRO, 2 , 57–75 .
Saaty, T. L. (1977). A scaling method for priorities in hierarchical structures. Journal of Mathematical Psychology, 15 , 234–281 .
Stewart, T. J. , & Janssen, R. (2013). Integrated value function construction with appli- cation to impact assessments. International Transactions in Operational Research, 20 , 559–578 .
Tsai, K.-H. , & Wang, J.-C. (2005). Does R&D performance decline with firm size?—A re-examination in terms of elasticity. Research Policy, 34 , 966–976 .
Tzeng, G.-H. , Lin, C.-W. , & Opricovic, S. (2005). Multi-criteria analysis of alternative– fuel buses for public transportation. Energy Policy, 33 , 1373–1383 .
Wilson, T. L. , & Anell, B. I. (1999). Business service firms and market share. Journal of Small Business Strategy, 10 , 41–53 .
Wulf, D. , & Bertsch, V. (2017). A natural language generation approach to sup- port understanding and traceability of multi-dimensional preferential sensitivity analysis in multi-criteria decision making. Expert Systems with Applications, 83 , 131–144 .
Xia, W. , & Wu, Z. (2007). Supplier selection with multiple criteria in volume dis- count environments. Omega, 35 , 494–504 .
Yager, R. R. (1988). On ordered weighted averaging aggregation operators in multi- criteria decision making. IEEE Transactions on Systems, Man, and Cybernetics, 18 , 183–190 .
Jafar Rezaei is an associate professor of operations and supply chain management at Delft University of Technology, the Netherlands, where he also obtained his Ph.D. One of his main research interests is in the area of multi-criteria decision-making (MCDM) analysis. He has published in various academic journals, including International Journal of Production Economics, International Journal of Production Research, Industrial Marketing Management, Applied Soft Computing, Applied Mathematical Modelling, IEEE Transactions on Engineering Management, Journal of Cleaner Production, European Journal of Operational Research, Information Science, Omega, and Expert Systems with Applications.