Jou nal o Indus ial Enginee ing and Managemen
JIEM, 2013 – 6(3): 796-804 – Online ISSN: 2013-0953 – P in ISSN: 2013-8423
h p://dx.doi.o g/10.3926/jiem.551
In eg a ing compu e -aided modeling and mic o-simula ion
in mul i-c i e ia e alua ion o se ice in as uc u e
assignmen app oaches
Al onso Du an, Isabel Ga cia, Miguel Gu ie ez
Uni . Ca los III de Mad id (Spain)
du [email protected], iga [email protected], mg e [email protected]
Recei ed: Sep embe 2012
Accep ed: Ma ch 2013
Abs ac :
Pu pose:
This pape p oposes an in eg a ed compu e -suppo ed mul i-s aged app oach o he
lexible design and mul ic i e ia e alua ion o se ice in as uc u e assignmen p ocesses/
algo i hms.
Design/me hodology/app oach:
I in ol es pa icula izing a me amodel encompassing he main
gene ic componen s and ela ionships in o p ocess models and p ocess ins ances, by
inco po a ing s uc u al da a om he eal-li e sys em. Exis ing da a on he a ge use
popula ion is ed in o a mic o-modeling sys em o gene a e a ma ching popula ion o indi idual
“ i ual” use s, each wi h i s own se o ai alues. The mic o-simula ion o hei in e ac ion
wi h he assignmen p ocess o bo h he incumben and he compe i o s gene a es a ich mul i-
dimensional ou pu , encompassing bo h “ e ealed” and non-obse able da a. This enables a
comp ehensi e mul i-c i e ia e alua ion o he o eseeable pe o mance o he designed
p ocess/ algo i hm, and he e o e i s i e a i e imp o emen .
Findings:
The esea ch p ojec de eloped a se o me hodologies and associa ed suppo ing ools
encompassing he modeling, mic o-simula ion and pe o mance assessmen o se ice
in as uc u e assignmen p ocesses.
-796-
Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
O iginali y/ alue:
The p oposed app oach acili a es, in a mul ic i e ia en i onmen , he lexible
modeling/design o si ua ion-speci ic assignmen p ocesses/algo i hms and hei pe o mance
assessmen when acing hei case-speci ic use popula ion.
Keywo ds:
in as uc u e assignmen , me amodeling, mic osimula ion, mul ic i e ia, simula ion pla o ms
1. In oduc ion and Objec i es
The assignmen o usage slo s o ixed capaci y, non-s o able se ice in as uc u es (such as
nigh s a a ho el ooms o u iliza ion o a su ge y ope a ing able o medical eme gency
esou ce) has bo h signi ican economic ele ance and me hodological complexi y. Faced wi h
he e ogeneous po en ial use s, he ealiza ion o he in as uc u e’s po en ial alue depends
on i s ac ual u iliza ion and on he app op ia eness o he ma ch wi h he ac ual use s i is
assigned o. Unused slo s c ea e no alue, and he alue o assigned slo s migh be e y
di e en depending on he speci ic assignmen , as illus a ed by he widely di e en p ices
o en paid o con iguous sea s in he same plane.
These alloca ion decisions, cha ac e ized by negligible a iable ( olume- ela ed) cos s in
compa ison wi h ixed cos s, “pe ishable” esou ces and he e ogeneous po en ial cus ome s
can gain insigh s om he applica ion o “Re enue Managemen ” (RM) app oaches (Chiang,
Chen & Xu, 2007; Abdelghany & Abdelghany, 2008). Gi en essen ially ixed cos and adop ing
p o i maximiza ion as he sole objec i e, RM ocuses on ex ac ing he maximum e enue
om a s eam o po en ial cus ome s, wi h di e en p ice sensi i i y, wi hin exis ing esou ce
cons ain s. Thus, RM applies p ice disc imina ion, selling essen ially he same se ice o
cus ome s in di e en “ a e classes” a di e en p ices.
RM app oaches, howe e , ha e wo majo limi a ions. Ad-hoc de elopmen o adap a ion o
assignmen algo i hms is equi ed o ailo hem o each o ganiza ion’s speci ic s uc u al
si ua ion (assignmen p ocess design,...). This migh be cumbe some and esul s in un ied
p ocedu es, pa icula ly gi en he pa icula i ies o each o ganiza ion’s use popula ion.
Addi ionally, sole ocus on di ec , sho - e m e enue/ p o i maximiza ion is o en excessi ely
na ow.
This pape p esen s an in eg a ed app oach o ackling hese issues. Thus, he majo
objec i es o he app oach discussed he e a e:
•Suppo he lexible modeling/design o si ua ion-speci ic assignmen
p ocesses/algo i hms by inco po a ing speci ic s uc u al da a o a gene alized model
(“me amodel”) o in as uc u e assignmen .
-797-
Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
•Facili a e c edible assessmen o he pe o mance o he esul ing designs aking in o
accoun he peculia i ies o ha si ua ion’s use popula ion.
•Enable app op ia e mul ic i e ia analysis.
This app oach and he enabling ools ha e been de eloped wi hin a 4-yea publicly unded
esea ch p ojec wi h he suppo o ou collabo a ing o ganiza ions.
2. S uc u e o he In eg a ed App oach
Figu e 1 depic s he gene al s uc u e o he in eg a ed app oach p oposed in his pape . This
sec ion con ains a gene al desc ip ion o he app oach, i s componen s, ela ions and suppo
ools, which a e hen discussed in subsequen sec ions.
Fo in as uc u e assignmen decisions, in cohe ence wi h he i s objec i e, a gene alized
me amodel has been de eloped, along wi h a me hodology suppo ed by a p oo -o -concep
compu e ool o enable he lexible modeling/design o case-speci ic assignmen
p ocesses/algo i hms by inco po a ing he s uc u al da a om he eal li e sys em’s
in as uc u e assignmen business p ocess. This componen is u he discussed in he
ollowing sec ion.
Figu e 1. S uc u e o he p oposed in eg a ed app oach
In o de o ake in o accoun he peculia i ies o each si ua ion’s use popula ion, all exis ing
cus ome da a o he eal sys em’s a ge segmen is inco po a ed in o a lexible cus ome
-798-
Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
mic o-modeling sys em, leading o he gene a ion o a popula ion composed o indi idual
“ i ual” cus ome s, each wi h i s own se o ai alues, ha closely esembles he eal
sys em’s use popula ion.
These indi idualized use s a e loaded in o an ad-hoc mic o-simula ion pla o m, whe e hey
ace bo h he assignmen business p ocess/algo i hm designed o he incumben o ganiza ion
and a simpli ied ep esen a ion o he subs i u i e p oduc s/compe i o ’s o e ings, as discussed
in sec ion 4.
The mic o-simula ion o he in e ac ion be ween he use popula ion, he incumben ’s
assignmen p ocess and ha o he compe i o s gene a es a ich mul idimensional ou pu ,
encompassing bo h “ e ealed” da a, ha would be obse able in eal li e, and non-obse able
da a. This enables a comp ehensi e mul i-c i e ia e alua ion o he o eseeable pe o mance o
he designed business p ocess/ algo i hm, ha can be used o i s i e a i e imp o emen , as
discussed in sec ion 5.
3. Business p ocess/assignmen algo i hms lexible modeling
As ou lined in he p e ious sec ion, ollowing he cha ac e iza ion o he in as uc u e
assignmen business p ocesses ha a e he ocus o his esea ch, hese a ge p ocesses ha e
been analyzed o syn hesize he gene ic model ha encompasses hem. This gene ic model o
business p ocess models, o business p ocess me amodel, includes he main gene ic
componen s o he in as uc u e assignmen models as well as hei ela ionships (OMG,
2006).
Figu e 2 depic s he UML class diag am o he p oposed me amodel (Gu ie ez & Du an, 2011).
An in as uc u e assignmen p ocess model de ines how a pa icula ype o in as uc u e is
assigned o one o i s po en ial uses. Cus ome s g ouped by cus ome segmen s (Cus ome
Segmen Type) demand he use o speci ic in as uc u es (In as uc u e Type) h ough
de ined access channels (Channel Type). In he mos gene al case, all Cus ome Segmen
Types would be able o access all In as uc u e Types h ough all Channel Types; he
exis ence o es ic ions in he access o some Cus ome Segmen Types, o access o ce ain
in as uc u es h ough some Channel Types, will lead o a comple e se o p oblems o be
add essed.
Each combina ion o possible ypes o in as uc u e access (In as uc u e Access Type) migh
be assigned di e en Value Types. In comme cial in as uc u e alloca ion p oblems, hese
Value Types will be he a i ypes, e.g. educed a i , no mal a i , p emium a i ; o e
p ice, lis p ice; high season, low season... The assignmen o a Value Type o an
In as uc u e Access Type de ines a gene ic Alloca ion Type. Each Alloca ion Type will apply
du ing a ime in e al, which will be de ined by a Gene ic S a Time and a Gene ic End Time.
These imes a e handled in he gene ic model laye in abs ac o m, wi hou aking a speci ic
-799-
Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
alue o ime. In some complex models logical ela ionships among hem (Cons ain s) will
ha e o be de ined. Acco ding o his app oach, a Business P ocess Model is de ined as a se o
Alloca ion Types.
Figu e 2. UML class diag am o he P ocess Me amodel
Each P ocess Model would gene ally yield o a numbe o P ocess Ins ances. Concep ually,
p ocess ins ances can be iden i ied as in as uc u e assignmen op imiza ion p oblems.
A compu e ool which has been buil upon he desc ibed me amodel assis s in s uc u ing eal
li e sys em da a in o de o de ine in as uc u e assignmen p ocess models and co esponding
p ocess ins ances h ough a guided sequence o windows-based o ms. I con ains a
ep esen a i e collec ion o in e nally de ined possible pa ame e and a iable ypes. The e is
also a p ocedu e ha au oma ically iden i ies o ecas ing equi emen s o hose a ibu es
s a ed as s a is ical pa ame e s, such as he mean o he dis ibu ion o he cus ome
segmen ’s a i al a e.
In a second se o windows-based o ms, he compu e ool helps he de ini ion o he p ocess
ins ance o be conside ed in a speci ic case, i.e., he assignmen business p ocess o be
designed o he incumben o ganiza ion. En i y a ibu es a e de ined o each ins ance,
speci ying i hey should be s a ed as a iables. These a ibu es a e loaded in he
co esponding da abase in an au oma ic p e-compiled p ocedu e. Las ly, he ool assis s in he
de ini ion o mul i-objec i e unc ions, h ough ma hema ical combina ions o a subse o
pa ame e s and a iables.
4. Mic o-modeling o he a ge use segmen and mic o-simula ion pla o m
In he mul i-s aged in eg a ed app oach p oposed in his pape , he ini ial assignmen
p ocess/algo i hm/ o ecas ing equi emen s designed wi h he compu e -aided me hodology
desc ibed in he p e ious sec ion analyze he use popula ion only a an agg ega ed le el, as
being composed o sub-segmen s cha ac e ized by a ibu es such as a i al dis ibu ion
pa ame e s. To accomplish he second objec i e (c edible pe o mance assessmen aking in o
-800-

Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
accoun he peculia i ies o each si ua ion’s use popula ion) and o enable he ul illmen o he
hi d (mul ic i e ia e alua ion), a compu e -suppo ed me hodology has been de eloped o he
mic o-modeling o he a ge use segmen and he co esponding mic o-simula ion o he
assignmen p ocess (Kle ma ken, 2002; Abdelghany & Abdelghany, 2008).
Thus, use s a e now modeled as indi idual “ i ual” so wa e agen s, each wi h i s own se o
ai alues. The me hodology calls o he in oduc ion, h ough se ies o menus and inpu
ables, o all a ailable da a on he cha ac e is ics o he a ge use s in he eal-li e sys em o
which he assessmen p ocess is being designed. This leads o he gene a ion o an
indi idualized ep esen a ion o he a ge use segmen in a way ha ully exploi s all exis ing
use in o ma ion o ensu e ha i esembles he eal sys em’s use popula ion as closely as
possible, in wo key dimensions: he in e nal co ela ion o he ai s (i.e., which ai s end o
happen oge he ) and he a e age alue dis ibu ion o each indi idual ai .
These indi idualized use s a e loaded in o an ad-hoc mic o-simula ion pla o m, whe e hey
ace bo h he assignmen business p ocess/algo i hm designed o he incumben o ganiza ion
and a simpli ied ep esen a ion o he subs i u i e p oduc s/compe i o ’s o e ings. The ma ch
be ween he ai s o each indi idual use and he cha ac e is ics o each se ice in as uc u e
al e na i e will in luence he a ac i eness o ha pa icula use o each al e na i e. As an
illus a ion o he pla o m’s design, Fig 3 shows a subse o he UML class diag am highligh ing
he gene a ion o his lexible “a ac i eness” o “u ili y unc ion”, named “Willingness To Pay”
in he pla o m bu ully gene ic in na u e.
The sys em’s beha io is hen simula ed as he agg ega ed e ec o hese indi idual decisions
o e ime. As discussed in he nex sec ion, his app oach, besides i s po en ial o a mo e
accu a e ep esen a ion o he sys em’s agg ega ed e olu ion, allows da a access, o analysis
and e alua ion pu poses, no only a a disagg ega ed le el bu also o da a dimensions ha
would be non-obse able in o he app oaches.
Figu e 3. Subse o UML class diag am highligh ing he gene a ion o he “Willingness o pay”
-801-
Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
5. Mul idimensional ou pu and mul i-c i e ia e alua ion
Mic o-simula ion p o ides a wide and ich se o ou pu a iables ha may be used o compa e
he pe o mance o al e na i e business p ocess designs, as well as o hei u he
e inemen . Modeling cus ome beha io ai s esul s in wo main ypes o ou pu a iables:
•Re ealed da a. This is he ype o da a ha could ac ually be egis e ed in a eal-li e
si ua ion, since i is obse able by he ho el, like oom occupa ion le els pe oom ype
and day, o daily e enue pe oom ype. In eal-li e cases, e en i his da a is
eco dable, i may be only pa ially eco ded, as he equi ed in o ma ion sys em and
p ocedu es may o may no be implemen ed.
•Non- e ealed da a. These a e ypically cus ome ai s, highly ele an o business
p ocess pe o mance, bu which may no be easily obse ed, e.g. he maximum p ice
ha a pa icula cus ome is eady o pay o a good/se ice (Willingness To Pay). I he
necessa y in o ma ion is a ailable, he simula ion pla o m assigns alues o hese
ai s o each indi idual use , hus p o iding a use ul ou pu o pe o mance
assessmen . Fo example, ho el cus ome s may be clus e ed in p ice ca ego ies
acco ding hei willingness o pay, and by compa ing hese igu es wi h he ac ual
assignmen pe o med, he e ec i eness o he assignmen p ocess may be assessed.
These wo se s o da a a e used o c ea e he dimensions o he mul i-c i e ia e alua ion o he
al e na i e designs conside ed. In some sec o s, such as heal h ca e, mul i-c i e ia e alua ion
has an immedia e i . Fu he mo e, e en in sec o s whe e p o i is he main pe o mance
c i e ion, a mul i-c i e ia app oach may also be use ul. Mid and long- e m e enues a e o en
dependen on aspec s such as cus ome loyal y o he company image, as p ojec ed h ough he
expe ience o cus ome s, ia e.g. commen s in web o ums. These elemen s, ha di ec ly o
indi ec ly impac e enue, can be o ganized along he ollowing e alua ion dimensions:
•Cu en e enue ob ained om he assigned in as uc u e.
•Fu u e e enue s eams, which may be subdi ided in o h ee componen s:
– Cus ome loyal y/ epe i ion. These indica o s a e ela ed o he epe i ion by he
same cus ome , and a e in luenced by such aspec s as he cus ome ’s pe cep ion o
he gap be ween se ice ecei ed and se ice expec ed.
– In luence on po en ial cus ome s. Indica o s ela ed o he in luence ha he
cus ome ’s pe cep ion may ha e on o he po en ial cus ome s, o ins ance h ough
commen s in oduced in o a web si e.
– Rela ed goods o se ices (c oss-selling). Indica o s ela ed o he likelihood ha
a ious ypes o cus ome s gene a e addi ional e enue h ough he consump ion o
-802-
Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
p oduc s o se ices ela ed o he assigned in as uc u e, e.g. expendi u es on he
ho el’s ancilla y se ices.
In sec o s whe e he objec i e unc ion is in insically mul i-c i e ia, e.g. heal h, in spi e o he
impo ance o economic c i e ia he e alua ion con ex is always mo e complex as social
aspec s play a c i ical ole, and a mul i-c i e ia app oach o e alua ion is likely o be he mos
app op ia e solu ion (Bana-Cos a, Fe nandes & Co eia, 2006).
This di e si y among sec o s in he assessmen c i e ia p e en s he p esc ip ion o a single
mul ic i e ia me hod. Ins ead, a lexible app oach is needed o choose he mos app op ia e
me hod o each case. Fo ins ance, in sec o s whe e he main e alua ion c i e ia a e
economic, a me hod like AHP may be sui able. This me hod is in ui i e o he use and
p o ides a e y good capabili y o c i e ia s uc u ing in successi e hie a chical le els. In he
case o sec o s mo e in insically mul ic i e ia, i may be mo e app op ia e o use a mo e
sophis ica ed non-compensa o y me hod, o ins ance, ELECTRE o PROMETHEE (Pome ol &
Ba ba-Rome o, 2000).
6. Conclusions
The in eg a ed compu e -suppo ed mul i-s aged app oach o he lexible design and
mul ic i e ia e alua ion o se ice in as uc u e assignmen p ocesses/ algo i hms p oposed in
his pape ackles he main limi a ions limi ing he applicabili y o Re enue Managemen
“con en ional” app oaches, namely hei equi emen o cumbe some, un ied ad-hoc
adap a ions o ca e o each o ganiza ion’s speci ic s uc u al si ua ion, and i s na ow ocus on
di ec , sho - e m p o i maximiza ion.
By way o he successi e pa icula iza ion o a gene alized me amodel in o p ocess models and
p ocess ins ances by inco po a ing s uc u al da a om he eal-li e sys em, i suppo s he
lexible modeling/design o si ua ion-speci ic assignmen p ocesses/algo i hms. By eeding
exis ing da a on he a ge use popula ion in o a mic o-modeling sys em o gene a e a
popula ion o indi idual “ i ual” use s, each wi h i s own se o ai alues, ha closely
esembles he eal sys em’s use s, i akes in o accoun i s peculia i ies. Th ough he ich
mul i-dimensional ou pu gene a ed by he mic o-simula ion o he in e ac ion o his
indi idualized popula ion wi h he assignmen p ocess o bo h he incumben and i s
compe i o s, i enables a comp ehensi e mul i-c i e ia e alua ion o he o eseeable
pe o mance o he designed p ocess/ algo i hm, and hus i s i e a i e imp o emen .
Acknowledgemen s
This wo k s ems om he pa icipa ion o he au ho s in a esea ch p ojec unded by he
Spanish Na ional Resea ch Plan, e e ence DPI2008-04872, i le "Op imizacion de la asignación
de in aes uc u as de se icios median e simulación - sec o es ho ele o y sani a io”
-803-
Jou nal o Indus ial Enginee ing and Managemen – h p://dx.doi.o g/10.3926/jiem.551
Re e ences
Abdelghany, A., Abdelghany, K. (2008). A mic o-simula ion app oach o Ai line Compe i ion
Analysis and Demand Modelling. In e na ional Jou nal o Re enue Managemen , 2(3),
287-306. h p://dx.doi.o g/10.1504/IJRM.2008.020626
Bana-Cos a, C.A., Fe nandes, T.G., Co eia, P.V.D. (2006). P io i isa ion o public in es men s in
social in as uc u es using mul ic i e ia alue analysis and decision con e encing: a case
s udy. In e na ional T ansac ions in Ope a ional Resea ch, 13(4), 279-297.
h p://dx.doi.o g/10.1111/j.1475-3995.2006.00549.x
Chiang, W.C., Chen, J.C.H., Xu, X. (2007). An o e iew o esea ch on e enue managemen :
Cu en issues and u u e esea ch. In e na ional Jou nal o Re enue Managemen , 1(1),
97-128. h p://dx.doi.o g/10.1504/IJRM.2007.011196
Gu ie ez, M., Du an, A. (2011). Gene ic Model Base Design o Decision Suppo Sys ems in
Re enue Managemen : Applica ions o Ho el and Heal h Ca e Sec o s. In C. Jao (ed.) E icien
Decision Suppo Sys ems: P ac ices and Challenges-F om Cu en o Fu u e, 183-194.
Kle ma ken, N.A. (2002). S a is ical in e ence in mic o-simula ion models: inco po a ing
ex e nal in o ma ion. Ma hema ics and compu e s in simula ion, 59(1.3), 255-265.
h p://dx.doi.o g/10.1016/S0378-4754(01)00413-X
OMG - Objec Managemen G oup (2006). Me a-Objec Facili y (MOF) V 2.0 co e speci ica ion.
h p://www.omg.o g/spec/MOF/2.0/PDF. Accessed 23 Feb 2012.
Pome ol, J.C., Ba ba-Rome o, S. (2000). Mul ic i e ion Decision in Managemen : P inciples and
P ac ice. Kluwe Academic Publishe s, No well.
Jou nal o Indus ial Enginee ing and Managemen , 2013 (www.jiem.o g)
A icle's con en s a e p o ided on a A ibu ion-Non Comme cial 3.0 C ea i e commons license. Reade s a e allowed o copy, dis ibu e
and communica e a icle's con en s, p o ided he au ho 's and Jou nal o Indus ial Enginee ing and Managemen 's names a e included.
I mus no be used o comme cial pu poses. To see he comple e license con en s, please isi
h p://c ea i ecommons.o g/licenses/by-nc/3.0/.
-804-