Au oma ic Se ice Ag eemen Nego ia o s in Open
Comme ce En i onmen s
Manuel Resinas, Pablo Fe nández, and Ra ael Co chuelo
ABSTRACT: The e is a s eady shi in e‑comme ce om goods o se ices ha mus be
p o isioned acco ding o se ice ag eemen s. This s udy ocuses on so wa e amewo ks
o de elop au oma ed nego ia o s in open comme ce en i onmen s. Analysis o he li e a‑
u e on au oma ed nego ia ion and ypical case s udies led o a ca alog o 16 objec i e
equi emen s and a concep ual model ha was used o compa e 11 s a e‑o ‑ he‑a so wa e
amewo ks. None o hem was well sui ed o nego ia ing se ice ag eemen s in open
comme ce en i onmen s. This mo i a ed wo k on a e e ence a chi ec u e ha p o ides
he ounda ions o de elop nego ia ion sys ems ha add ess he p e ious equi emen s.
A so wa e amewo k was de ised o alida e he p oposal by means o case s udies.
The s udy con ibu es o he ields o equi emen s enginee ing and so wa e design, and
is expec ed o suppo u u e e o s o p ac i ione s and esea che s because i s indings
b idge he gap among he exis ing au oma ed nego ia ion echniques and lay he ounda‑
ions o de eloping new so wa e amewo ks.
Key woRdS ANd phRASeS: Au oma ed nego ia ion, elec onic se ices, nego ia ion
equi emen s, e e ence a chi ec u e, se ice ag eemen s, se ices science, so wa e
amewo ks.
E-comme ce has e ol ed om shopping o goods on he In e ne o ou sou c-
ing elec onic se ices, such as ligh ese a ions, paymen s, o execu ing
business in elligence jobs on he cloud [10, 47, 56]. The ocus o he p esen
s udy is on so-called nex -gene a ion companies, which ely hea ily on he
au oma ion o business p ocesses ha build on se ices ou sou ced a ound he
wo ld because his enables companies o be mo e e icien and cos -e ec i e by
exploi ing economies o scale [4, 12, 13, 14, 15, 22, 40, 55]. Fo ins ance, Ga ne
p edic s ha a leas 30 pe cen o he in es men in so wa e will go o ou -
sou ced se ices ins ead o p oduc licenses by 2012, and 40 pe cen o capi al
expendi u es will be made o ou sou ced in as uc u e by 2011 [44].
Se ices science ocuses on me ging esul s in compu e science, so wa e
enginee ing, and classical business sciences so ha business equi emen s
a e be e mapped on o echnology [5]. The complex issues in ol ed in his
mapping equi e he simul aneous de elopmen bo h o business me hods
and o he echnology ha suppo s hem. Fo una ely, he echnologies o
se ice-o ien ed a chi ec u e (SOA) seem o be helping business analys s
This wo k was unded in pa by he Eu opean Commission
(FEDER—El Fondo Eu opeo de Desa ollo Regional), he Spanish Minis y o Sci-
ence and Inno a ion, and he Andalusian go e nmen . The wo k by Manuel Resinas
and Pablo Fe nández was suppo ed by g an s TIN2006–00472 (Web-Fac o ies),
TIN2009–07366 (SETI), and P07-TIC-2533 (Isabel); he wo k by Ra ael Co chuelo
was suppo ed by g an s TIN2007–64119, P07-TIC-02602, P08-TIC-4100, and
TIN2008–04718-E (In eg aWeb).
and so wa e a chi ec s b idge he gap be ween he wo wo lds
seamlessly [14, 42]. Acco ding o O acle, companies ha implemen an SOA
a e able o educe he cos s o he in eg a ion and main enance o p ojec s
by a leas 30 pe cen [58].
Howe e , his is no enough: I is also necessa y o be e ec i e, e icien ,
and lexible in he as -changing condi ions o he cu en ma ke .
Ou sou cing is expec ed o e ol e in o dynamic ou sou cing, whe eby
au oma ed business p ocesses can sea ch, assess, and selec he app op ia e
IT se ice p o ide s and le els on demand [13, 26]. Fo dynamic
ou sou cing o become a eali y, a numbe o c i ical issues mus be
add essed— o example, seman ic disco e y, se ice ag eemen nego ia ion,
in e ope abili y, moni o ing, managemen , and go e nance, o name a ew.
In he ield o se ice ag eemen nego ia ion, he de elopmen o
p o ocols, decision-making algo i hms, and p e e ences and ag eemen
models wi h desi able cha ac e is ics pose impo an challenges ha ha e
s imula ed sig-ni ican esea ch e o s by he au oma ed nego ia ion
esea ch communi y. Howe e , he success ul de elopmen o au oma ed
nego ia ion sys ems e-qui es an unde s anding o he whole sys em as well
as o he equi emen s o he nego ia ion con ex so ha he mos
app op ia e nego ia ion p o ocol, decision-making algo i hms, and model
a e selec ed and pu oge he in an au oma ed nego ia ion sys em [29].
This a icle akes a so wa e enginee ing app oach in dealing wi h he
la e p oblem. The ul ima e goal is o unde s and he equi emen s o
au oma ed nego ia ion sys ems used o make se ice ag eemen s in open
comme ce en i onmen s and o p o ide he ounda ions o de eloping
such sys ems. The con ibu ions in his a icle ocus on wo a eas o he
cu en so wa e enginee ing body o knowledge [1].
The i s con ibu ion is o he ield o equi emen s enginee ing. An
exhaus- i e e iew o he li e a u e oge he wi h analysis o se e al case
s udies led o he iden i ica ion o se e al key equi emen s in he con ex o
au oma ed nego ia ion o se ice ag eemen s in open comme ce
en i onmen s. No e ha acco ding o Bleis ein e al., li le a en ion has
been gi en o compiling a ca alog o equi emen s o e-business
applica ions [9]. The p esen e o , hen, can be seen as a ma e ializa ion o
such a ca alog o se ice ag eemen nego ia ion sys ems. The key
equi emen s a e ela ed o he exp essi eness o se ice ag eemen s, he
he e ogenei y o he pa ies in ol ed in a nego ia ion, he lack o comple e
in o ma ion abou hem, and he implici dynamism o open comme ce
en i onmen s. Building on hese ac o s, a concep ual model was
de eloped based on 16 objec i e c i e ia ha make i possible o compa e
cu en and u u e p oposals. This analysis leads o he conclusion ha he
nego ia ion con ex , which depends on he use o he au oma ed nego ia ion
sys em and he pa ies wi h which he sys em nego ia es, is likely o change
in open comme ce en i onmen s. Consequen ly, au oma ed nego ia ion sys-
ems o such en i onmen s should be able o accommoda e all hese changes
wi hou equi ing a g ea deal o e o .
The second con ibu ion pe ains o so wa e design. Since he e is an
impo an need o emb ace change, he nex logical s ep is o analyze he
cu en li e a u e on so wa e amewo ks o de elop au oma ed nego ia ion
sys ems. The conclusion is ha nei he p o ocol-o ien ed amewo ks no
in elligence-o ien ed amewo ks add ess he a o emen ioned key equi e-
men s simul aneously, which shows he concep ual complexi y o pu ing
oge he au oma ed nego ia ion echniques in a unique amewo k [3, 6, 7,
24, 30, 32, 35, 43, 45, 52, 54]. This mo i a ed he e o o de elop a e e ence
a chi ec u e o au oma ed nego ia ion sys ems. A e e ence a chi ec u e is no
he desc ip ion o a conc e e sys em, bu a eusable design ha de ines he
asks, g ouped in o oles, and he in e ac ions ha mus be implemen ed in an
au oma ed nego ia ion sys em. I is a ounda ion o de eloping au oma ed
nego ia ion sys ems ha acili a es hei main ainabili y and hei adap abili y
o changes in hei con ex .
Finally, he hi d con ibu ion is he so wa e amewo k ha has been
de eloped. I demons a es he soundness and use ulness o he e e ence
a chi ec u e, and, as well, p o ides a e e ence implemen a ion in which au-
oma ed nego ia ion echniques (p o ocols, algo i hms, and models) can be
in eg a ed. I was alida ed by means o se e al case s udies.
These con ibu ions can be used o suppo u u e e o s o p ac i ione s
and esea che s in he au oma ed nego ia ion ield.
P ac i ione s can use he equi emen s and he e e ence a chi ec u e o
se he basis on which new so wa e amewo ks o au oma ed nego ia ion
sys ems can be buil . No e, oo, ha a numbe o p o ocols, algo i hms, and
models ha e al eady been in eg a ed in o an accompanying so wa e oolki .
A p ac i ione o whom his oolki is enough can deploy he amewo k im-
media ely. Ou implemen a ion o he so wa e amewo k may be an excel-
len s a ing poin in de eloping ad anced ea u es, such as in eg a ion wi h
o he en e p ise sys ems (e.g., en e p ise se ice buses o business p ocess
managemen sys ems).
Resea che s will ob ain h ee old suppo om he e e ence a chi ec u e.
Fi s , he equi emen s iden i ied may help guide esea ch e o s on new
nego ia ion models ha deal be e wi h he p oblems iden i ied ega ding
au oma ed nego ia ions in open comme ce en i onmen s. Second, he a chi-
ec u e may help esea che s on au oma ed nego ia ion o be awa e o he
asks o which hei nego ia ion model mus accoun . Mo e impo an , he
a chi ec u e enables esea che s o ocus on a pa icula pa o an au oma ed
nego ia ion sys em (e.g., wo ld modeling algo i hms o esponse gene a ion
algo i hms). A he same ime i b idges he gap be ween cu en au oma ed
nego ia ion echniques (p o ocols, algo i hms, models), and hus eases he
euse o cu en esul s in his ield. Finally, since he e is an implemen ed
so wa e amewo k based on he e e ence a chi ec u e, a esea che who
wo ks on decision-making can in eg a e a p oposal in o he amewo k, es
i , and compa e he esul s. This enables compa ison o se e al p oposals
om an empi ical poin o iew, which has usually been a majo d awback
acco ding o he li e a u e su eyed.
Concep ual Model and Requi emen s o Au oma ed
Nego ia ion Sys ems
The discussion in his sec ion de ails he concep ual model o an au oma ed
nego ia ion sys em, hen analyzes he key p oblems a se ice ag eemen
nego ia ion sys em mus ace in open comme ce en i onmen s, and epo s
on 16 equi emen s o au oma ed nego ia ion sys ems o deal wi h hese
p oblems. Finally, i epo s on he p oposals su eyed and analyzes how
well hey suppo hese equi emen s.
Concep ual Model
Figu e 1 ske ches a concep ual map ega ding nego ia ion sys ems ha
de i es om he su ey o he li e a u e and analysis o se e al case s udies.
The use o an au oma ed nego ia ion sys em de ines p e e ences, namely, he
da a used o ensu e ha an ag eemen is eached acco ding o he use ’s
needs. The use ini ia es he au oma ed nego ia ion sys em o nego ia e on his
o he behal . The au oma ed nego ia ion sys em is also p o ided wi h pa y
e e ences. A pa y e e ence gi es a means o in e ac wi h a pa y. The use may
p o ide pa y e e ences o hey may come om a pa y as a eques o s a a
nego ia ion.
Fo each pa y e e ence ecei ed, he au oma ed nego ia ion sys em s a s a
p ocess whose goal is o execu e a nego ia ion wi h he pa y whose e e ence was
ecei ed. a nego ia ion is a conc e e execu ion o a nego ia ion p o ocol played
by wo o mo e pa ies. nego ia ions a e con igu ed by a p o ocol con igu a ion,
which speci ies cha ac e is ics o he nego ia ion p o ocol o a pa icula execu-
ion. Fo ins ance, hey may speci y he imeou o he p o ocol o i s secu i y
ea u es.
The execu ion o a nego ia ion in ol es he exchange o nego ia ion messages,
whose speci ic cha ac e is ics a e de e mined by a nego ia ion p o ocol. Each
nego ia ion message has a pe o ma i e o exp ess he in en ion o he sende
abou he message and a message con en . Depending on he pe o ma i e, nego-
ia ion messages can be classi ied as binding nego ia ion messages, which in ol e
a i m commi men wi h he o he pa y, and nonbinding nego ia ion messages,
which do no in ol e a i m commi men . nonbinding nego ia ion messages can
be used o gi e addi ional in o ma ion o he o he pa y o o explo e new
choices wi hou commi ing o hem. This is e y use ul in dynamic con ex s
in which he e a e se e al simul aneous nego ia ions wi h di e en pa ies and
he easibili y o commi ing o an ag eemen depends on he cu en s a e o
he se ice p o ide ’s esou ces.
The message con en o a nego ia ion message is usually a p oposal. Howe e ,
o he kinds o in o ma ion can be exchanged, such as h ea s, ewa ds, o a gu-
men s. A p oposal is an o e o an ag eemen made by one pa y. I speci ies he
pa ies o which i pe ains and a se o e ms. e ms designa e bo h unc ional
desc ip ions and non unc ional gua an ees o he se ice. Some examples o
usual e ms include: “ he se ice in e ace is speci ied in he documen ha is
a ailable a h p://example.o g/myse ice.wsdl,” “ he esponse ime is less
han 20 ms,” o “ he numbe o se ice eques s is lowe han 10 imes pe
minu e.” Addi ionally, p oposals may also include nego ia ion da a o exp ess
addi ional in o ma ion o guide he nego ia ion p ocess.
When pa ies ag ee on a p oposal, an ag eemen is c ea ed. An ag eemen is a
documen ha de ines a ela ionship be ween pa ies. I s goal is o de ine he
e ms ha egula e he execu ion o a se ice, and i mus ha e a speci ica ion
o he pa ies in ol ed and a collec ion o e ms, such as hose desc ibed in
he p oposal. These e ms egula e how he execu ion o he se ice mus be
ca ied ou in he con ex o he ag eemen . In addi ion, unlike p oposals, in
which he e ms can be le open in o de o be e ined la e , ag eemen e ms
mus be ully speci ied and ambigui ies mus be a oided.
The execu ion o a nego ia ion equi es he au oma ed nego ia ion sys em o
make se e al decisions on he c ea ion o nego ia ion messages (wha nego ia ion
message mus be sen o he o he pa y), he commi men o binding nego ia ion
messages (whe he he au oma ed nego ia ion sys em mus commi o an ag ee-
men and when i mus do i ), and he decommi men om p e iously c ea ed
ag eemen s. These decisions a e suppo ed by he in o ma ion a ailable o he
au oma ed nego ia ion sys em, which is he se o p e e ences sen by he use ,
a wo ld model he au oma ed nego ia ion sys em may main ain, and ex e nal
ac o s ha may ha e an in luence on he decision, such as he p o ide ’s
capaci y o accep a new ag eemen .
Finally, he wo ld model main ained by he au oma ed nego ia ion sys em may
comp ise models c ea ed abou he ma ke , o he nego ia ing pa ies, o he ne-
go ia ion domain. To c ea e hese models, he au oma ed nego ia ion sys em needs
o ga he in o ma ion om se e al sou ces, which include domain expe s, public
in o ma ion abou p e e ences acili a ed by he o he pa ies hemsel es, ex e nal
in o ma ion p o ide s, and by means o an analysis o p e ious nego ia ions.
Figu e 1. Concep ual Map o Au oma ed Nego ia ion Sys ems

Key P oblems and Requi emen s
Exp essi eness o Se ice Ag eemen s
The nego ia ion o a se ice ag eemen usually in ol es such e ms as a ail-
abili y, esponse ime, secu i y, and p ice. The nego ia ing pa ies a e able o
make ade-o s among he e ms acco ding o hei p e e ences. The e o e,
o an au oma ed nego ia ion sys em o suppo exp essi e-enough
se ice ag eemen s, i should:
1.1. Suppo mul i e m nego ia ion p o ocols. No all nego ia ion
p o ocols allow o mul i e m se ice ag eemen s. Fo ins ance,
mos auc ioning p o ocols, excep o mul ia ibu e auc ions,
only suppo he nego ia ion o one e m, usually he p ice [8, 28].
Ba gaining p o ocols, which in ol e exchanging p oposals and
coun e p oposals among he pa ies, usually suppo mul i e m
nego ia ions.
1.2. Manage exp essi e p e e ences. To enable ade-o s, p e e ences
mus be exp essed in a o malism ha cap u es he ela ionships
be ween e ms, o example, u ili y unc ions, combina ions o a -
ibu es, o uzzy cons ain s [17, 19, 37].
Pa ies A e He e ogeneous
The bes nego ia ion p o ocol, decision-making algo i hm, p e e ences model,
and ag eemen model depend on he nego ia ion con ex , which in ol es a
leas wo s akeholde s: he use o he au oma ed nego ia ion sys ems and
he pa ies wi h which he sys em nego ia es [29]. The p oblem is ha in open
comme ce en i onmen s, his nego ia ion con ex is likely o change. These
changes de i e om wo sou ces.
Fi s , in open comme ce en i onmen s, new pa ies may appea unex-
pec edly, implemen a a ie y o nego ia ion p o ocols, and ha e di e ging
beha io s du ing he nego ia ion. Fo example, some pa ies can concede
mo e a he beginning o he nego ia ion, whe eas o he s may concede only
when he deadline is app oaching and exp ess p e e ences and ag eemen s
using di e en models.
Second, he equi emen s o he use o an au oma ed nego ia ion sys em
ha e a s ong in luence on he sys em because he e is a ade-o be ween
he exp essi eness o he ag eemen and p e e ences models and he a ail-
abili y and complexi y o he co esponding decision-making algo i hms.
The e o e, models and algo i hms may change depending on he use ’s
needs. The p oblem he e is ha he use ’s needs a e subjec o con inuous
adap a ion and a ia ion, adding new business ules and egula ions, ypes
o business- ela ed e en s, ope a ions, and so o h [41].
Au oma ed nego ia ion sys ems in open comme ce en i onmen s should
be able o accommoda e all hese changes wi hou equi ing a g ea de el-
opmen e o . Hence, i would be desi able o an au oma ed nego ia ion
sys em o:
2.1. Suppo mul iple nego ia ion p o ocols. Since he e is no s anda d
nego ia ion p o ocol, di e en pa ies may implemen di e en
nego ia ion p o ocols. An au oma ed nego ia ion sys em should
suppo se e al nego ia ion p o ocols o a oid losing business
oppo uni ies.
2.2. Nego ia e he nego ia ion p o ocol. Al hough all ba gaining
p o ocols in ol e he exchange o p oposals be ween pa ies, he
exchange may be ca ied ou wi hou es ic ions on he con en o
he p oposal and coun e p oposal o wi h es ic ions on he o de
in which e ms a e nego ia ed o on he e ms o he p oposals [18,
20, 31]. I is desi able o nego ia ion sys ems o be able o choose
he mos sui able nego ia ion p o ocol o each con ex —depend-
ing, o example, on he numbe o e ms unde nego ia ion o he
nego ia ion deadline.
2.3. Suppo mul iple decision-making algo i hms. The e a e a
a ie y o decision-making algo i hms based on game- heo e ic ap-
p oaches, heu is ic app oaches, and e olu iona y app oaches [18,
19, 21, 33, 37, 38]. Thei e ec i eness depends on he beha io o
he o he pa ies [29, 46]. An au oma ed nego ia ion sys em should
suppo se e al decision-making algo i hms and choose he mos
app op ia e one a un ime.
2.4. Suppo mul iple ag eemen models. Ag eemen s can be ex-
p essed using o malisms ha ange om name- alue pai s o
on ologies o deon ic logic [6, 18, 25]. Suppo ing mul iple ag ee-
men models enables he use o make a ade-o be ween he
exp essi eness i equi es and he a ailabili y and complexi y o
he co esponding decision-making algo i hms.
2.5. Suppo mul iple p e e ences models. The e a e many o mal-
isms o exp ess p e e ences: u ili y unc ions, cons ain s, uzzy
cons ain s, a combina ion o a ibu es, and ules [16, 17, 18, 23,
24, 33, 37]. Two issues mus be conside ed be o e deciding which
o malism is he mos app op ia e: he nego ia ion domain and he
in luence o he model on he decision-making algo i hms wi h
ega d o a ailabili y and complexi y.
2.6. Allow o use p e e ences abou nego ia ion p ocesses. Usu-
ally, no all pa ies ha e he same p e e ences abou a nego ia ion
p ocess. Some pa ies may ha e a sho e deadline, may be eage
o each an ag eemen , o may be less s ic abou he ag eemen s
hey accep . Fu he mo e, he use ha se s he p e e ences may
be a so wa e sys em. Fo ins ance, a componen may analyze he
cu en s a e o he business p ocesses and p o ide he au oma ed
nego ia ion sys em wi h he app op ia e guidelines o he ne-
go ia ion depending on he cha ac e is ics and he s a e o hose
p ocesses (e.g., i a business ask is on a slack pa h o a p ocess
wo k low). This enables he in eg a ion o he au oma ed nego ia-
ion sys em wi h o he pa s o he IT in as uc u e.
Pa ial In o ma ion Abou Pa ies
Ha ing in o ma ion abou o he pa ies s eng hens one’s nego ia ing capabil-
i y o , complemen a ily, weakens o he s’ capabili ies. Un o una ely, au oma ed
nego ia ion sys ems do no usually ha e comple e in o ma ion abou he pa -
ies wi h which hey nego ia e [11, 37, 59]. The e o e, i would be desi able o
an au oma ed nego ia ion sys em o:
3.1. Manage di e en ypes o knowledge abou o he pa ies. This
can be ei he knowledge abou hei p e e ences, knowledge abou
hei beha io du ing he nego ia ion (e.g., whe he hey end o
concede, hei nego ia ion deadline), and knowledge abou he
pa y i sel (e.g., epu a ion o geog aphical loca ion) [39, 59].
3.2. Ga he in o ma ion om di e en sou ces. This includes in o -
ma ion p o ided by a domain expe . Fa a in, Sie a, and Jen-
nings p opose ha a domain expe mus p o ide a measu e o
simila i y be ween alues o e ms in he nego ia ion domain [19].
The e is also in o ma ion ga he ed di ec ly om he o he pa y.
Fo example, in WS-Ag eemen , pa ies’ empla es can be used o
lea n wha kind o ag eemen s o he pa ies a e willing o accep
[2]. In o ma ion may also be ga he ed om ex e nal in o ma ion
p o ide s, such as epu a ion p o ide s.
3.3. Build analysis-based models o pa ies. Messages exchanged wi h
o he pa ies du ing p e ious nego ia ions can be analyzed o lea n
abou hei p e e ences and hei beha io [11, 59]. This analysis
can be classi ied in o on-line analysis and o -line analysis, de-
pending on whe he he s a e o cu en nego ia ions is aken in o
accoun [34, 50, 59]. Some p oposals use bo h kinds o analysis; o
example, he one by Coehoo n and Jennings [11].
Ma ke s A e Dynamic
The idleness o a esou ce esul s in a loss o e enue [24]. In consequence,
p o ide s commonly o e discoun p ices when hei esou ces a e likely o
become idle. Se e al p o ide s and consume s usually compe e on he same
se ices, which makes ma ke condi ions ex emely ola ile. To deal wi h hese
changing ma ke condi ions, an au oma ed nego ia ion sys em should:
4.1. Suppo se e al nego ia ions simul aneously. This is desi able
because i would allow he sys em o choose he pa y ha o e s
he mos p o i able ag eemen .
4.2. Selec decision-making algo i hms dynamically. When dealing
wi h simul aneous nego ia ions, he s a e o he nego ia ions can
ha e an in luence on he decision-making algo i hms used. Fo
ins ance, i he sys em is ca ying ou se e al simul aneous nego-
ia ions and inds a e y p o i able ag eemen , i can ake a oughe
s ance wi h he o he pa ies. The e o e, i is desi able o be able o
change he decision-making algo i hm a un ime, so ha i can be
adap ed o changing con ex s e ec i ely [46].
4.3. Suppo decommi men . Decommi ing om an ag eemen in-
ol es e oking i and paying a decommi ee [39, 48]. In a dynam-
ic ma ke , mo e p o i able new o e s may be ound a any ime.
Hence, i is e y con enien o be able o decommi om p e ious
ag eemen s. This opic is no su icien ly co e ed in he li e a u e
and equi es u he esea ch be o e a comple e amewo k can be
de eloped.
4.4. Supe ised c ea ion o ag eemen s. To a oid commi ing o ag ee-
men s ha canno be sa is ied, he au oma ed nego ia ion sys em
should be supe ised by ex e nal elemen s, such as a capaci y
es ima o o de e mine whe he an ag eemen can be accep ed o
no [36].
4.5. Build ma ke models. The cha ac e is ics o he ma ke may ha e
an in luence on he nego ia ion p ocess [50]. The e o e, i is con-
enien o an au oma ed nego ia ion sys em o build models o
he ma ke o ob ain in o ma ion, such as he ese a ion p ice o
a p oduc o he chances ha new pa ies will be ound du ing a
nego ia ion [34].
Analysis o Cu en Solu ions
The key p oblems desc ibed in he p e ious sec ion p o ide a numbe o
objec i e equi emen s ha can be used o compa e cu en s a e-o - he-a
au oma ed nego ia ion amewo ks. Table 1 and Table 2 summa ize he com-
pa ison: a  in a cell means ha he co esponding p oposal p o ides explici
suppo o he co esponding ea u e; a ~ indica es ha i add esses i pa -
ially; an  indica es ha he ea u e is no suppo ed; NA means ha he e
is no in o ma ion a ailable. (No e ha p oposals such as hose o Kowalczyk
and o Su e al. a e no aken in o accoun because hey a e speci ic-pu pose
nego ia ion sys ems, no so wa e amewo ks [33, 53]. They a e no in ended
o be he ounda ion o building o he nego ia ion sys ems, which is he ocus
in his a icle.)
P o ocol-O ien ed F amewo ks
These amewo ks p o ide he b awn o a nego ia ion sys em because hey
deal wi h he nego ia ion p o ocol and low-le el in e ope abili y issues.
Some o hem de ine a nego ia ion hos o ma ke place ha ac s as a me-
dia o among he nego ia ing pa ies. Fo ins ance, Kim and Sege desc ibe a
Web se ices–enabled ma ke place a chi ec u e, which enables he au oma ion
he p o ide ’s esou ces o analyze whe he i has enough capaci y o p o i-
sion a p oposal (REQ 4.4).
The sepa a ion o he decision-making in o hese oles enhances he eus-
abili y o he di e en decision-making algo i hms and makes i easie o adap
he sys em o changes in i s nego ia ion con ex (REQ 2.3).
Wo ld Modeling Module
The Wo ld Modeling module ga he s, analyzes, and manages use ul in o ma-
ion wi h which o make decisions du ing a nego ia ion. I is composed o an
Inqui e , an In o man , and se e al Wo ldModelle s. Inqui e and In o man enable
polling he o he pa ies o ge in o ma ion abou hem and he cha ac e is ics
o he se ice demanded o o e ed. Inqui e ga he s in o ma ion om o he
pa ies by polling hei In o man s. The in o ma ion ga he ed om o he ne-
go ia ing pa ies is s o ed in esou ce Pa yCon ex da a.
An au oma ed nego ia ion sys em may ha e se e al Wo ldModelle s ha
can be g ouped in o wo ca ego ies: Wo ldModelle s de eloped by domain
expe s ha build models o he nego ia ion domain (e.g., a simila i y measu e
be ween alues o e ms in he nego ia ion domain [19]), and Wo ldModelle s
ha ga he in o ma ion om Ex e nalIn o ma ionP o ide s and messages
exchanged in nego ia ions (REQ 3.2) and analyze hem o build models o
pa ies o he ma ke (REQ 4.5) [11, 34, 46, 50, 59]. No e ha i is impo an
o ma ke modele s o ha e access o Ex e nalIn o ma ionP o ide s o ga he
ma ke in o ma ion, such as que ying a se ice egis y o ob ain in o ma ion
abou he numbe o p ospec i e p o ide s (o compe i o s) in he ma ke ,
o que ying se e al auc ion si es o ga he in o ma ion abou he esul s o
ecen auc ions o p o ide measu es, such as ecommended maximum p ices
o an auc ion [27, 50].
To upda e hei models, Wo ldModelle s can pe o m o -line analysis o p e-
ious in e ac ions using he nego ia ionhis o y o on-line analysis by means o
a publish/subsc ibe mechanism p o ided by he en i onmen al esou ces ha
no i ies hem when an e en ele an o hei models akes place (REQ 3.3).
No e ha his design based on independen Wo ldModelle s p o ides a
lexible s uc u e wi h which o manage di e en ypes o knowledge abou
pa ies easily (REQ 3.1), since each Wo ldModelle can ocus on one ype o
knowledge abou pa ies.
Coo dina ion Module
The Coo dina ion module coo dina es he h ee le els o coo dina ion con ex s
de ined in NegoFAST-Co e (sys em con ex , pa y con ex , nego ia ion con ex )
by means o he h ee oles o which his module is composed.
sys emCoo dina o coo dina es he in e ac ions be ween he sys em and he
use , he ini ializa ion and e mina ion o he sys em, and he ecep ion o
pa y e e ences om he use , he P o ocolnego ia o , and he P o ocolhandle .
The e e ences a e sen o he Pa yCoo dina o o be p ocessed. I also upda es
he sys emCon ex da a.

Pa yCoo dina o manages he p ocessing o a pa y e e ence be o e he
nego ia ion o ge in o ma ion abou he pa y ia an Inqui e , decide he nego-
ia ion p o ocol by means o a P o ocolnego ia o , and delega e he nego ia ion
i sel o a nego ia ionCoo dina o . I also upda es he Pa yCon ex da a.
nego ia ionCoo dina o coo dina es he execu ion o a nego ia ion by ac -
ing as a b idge be ween he P o ocolhandle and he ResponseGene a o . I
in okes he Commi handle when an app o al o send a binding nego ia-
ion message is necessa y. I s o es he s a us o he nego ia ion in esou ce
nego ia ionCon ex da a.
En i onmen al Resou ces
The en i onmen al esou ces in he NegoFAST-Co e e e ence a chi ec u e a e
da a s o es ha can be ead, modi ied, o bo h by he o he oles. In addi ion,
hey p o ide a publish/subsc ibe mechanism o no i y he oles when e en s
ake place. En i onmen al esou ces can be g ouped in o esou ces ha a e
eini ialized in each execu ion o he sys em and esou ces ha keep hei
in o ma ion be ween di e en execu ions.
The o me a e he ag eemen sResou ce, which s o es all o he ag eemen s
made by he sys em so ha hey can be analyzed on he ly— o example, o de-
cide whe he i is con enien o decommi om an ag eemen . The P e e ences-
Resou ce s o es he use s’ p e e ences and allows he o he oles o e alua e
and compa e ag eemen s and p oposals. sys emCon ex da a, Pa yCon ex da a,
and nego ia ionCon ex da a s o e in o ma ion ega ding he whole au oma ed
nego ia ion sys em (e.g., he ime when he sys em was ini ialized, known
pa y e e ences), each nego ia ing pa y (e.g., he in o ma ion ga he ed by he
Inqui e , he nego ia ion p o ocol selec ed, he esul o he nego ia ion), and
each nego ia ion (e.g., i s cu en s a e, he nego ia ion messages exchanged
wi h he o he pa ies).
The la e a e he Wo ldModel, which s o es he knowledge gene a ed by
he Wo ldModelle s, and he nego ia ionhis o y, which allows i o build models
based on p e ious in e ac ions.
NegoFAST-Ba gaining
The NegoFAST-Ba gaining e e ence a chi ec u e ex ends NegoFAST-Co e
o deal wi h he speci ic equi emen s o concu en ba gaining nego ia ions
(see Figu e 4). A sequence diag am o a ypical in e ac ion be ween he oles
in NegoFAST-Ba gaining is depic ed in Figu e 5.
P o ocolHandle
The P o ocolhandle mus adap ba gaining p o ocols in o a gene ic nego ia ion
p o ocol wi h he ollowing cha ac e is ics: i is bila e al (i.e., he e is only an
ini ia o and a esponde ), i is sequen ial ( he same pa y canno send wo
nego ia ion messages in a ow excep o wi hd aw and cancella ion messages),
and i is p oposal-based ( he message con en is composed o p oposals only.
In NegoFAST-Ba gaining, he P o ocolhandle is e ined in o a speci ic-pu pose
ole called Ba gainingP o ocolhandle .
Nego ia ionCoo dina o
The nego ia ionCoo dina o mus be designed o suppo concu en bila e al
nego ia ions (REQ 4.1). To his end, i is e ined in o Ba gaining Coo dina o ,
Bila e al nego ia o , and PoliciesManage . Ba gainingCoo dina o o ches-
a es he Bila e alnego ia o , he Commi handle , and he Pa yCoo dina-
o , and s o es he cu en s a e o he concu en nego ia ions in esou ce
Ba gainingCon ex da a.
Bila e alnego ia o ca ies ou a single bila e al nego ia ion by o ches a -
ing he Ba gainingP o ocolhandle and he Pe o ma i eselec o . I communi-
ca es wi h he Ba gainingCoo dina o o ask o app o al be o e sending a
binding nego ia ion message, and i ecei es nego ia ion policies om he
PoliciesManage .
PoliciesManage uses he nego ia ion guidelines p o ided by he use ’s
p e e ences oge he wi h he cu en s a e o he nego ia ions o de e mine spe-
ci ic nego ia ion policies ha will guide he beha io o he ResponseGene a o
(i.e., whe he i should concede in he nex p oposal and how much i should
Figu e 4. NegoFAST-Ba gaining Re e ence A chi ec u e
Figu e 5. NegoFAST-Ba gaining Re e ence A chi ec u e (sequence diag am)
concede) du ing he nego ia ion and sends hem o he Bila e alnego ia o s. By
means o he nego ia ion policies, he PoliciesManage may guide he beha io
o one nego ia ion based on how well he o he nego ia ions a e pe o ming
and, hence, p ope ly suppo concu en bila e al nego ia ions (REQ 4.1). Fo
ins ance, i one nego ia ion is pe o ming pa icula ly well, (i.e., he p oposals
om he o he pa y a e e y appealing), he nego ia ion policies o he o he
concu en nego ia ions can be se o make he ResponseGene a o concede
less. These nego ia ion policies a e also used o ensu e ha he conduc o
nego ia ions complies wi h he nego ia ion guidelines p o ided in he use ’s
p e e ences (REQ 2.6).
Response Gene a o
Since nego ia ion messages a e composed o a pe o ma i e and a p oposal, he
ResponseGene a o can be e ined in o he Pe o ma i eselec o , which selec s he
pe o ma i e o be used, and he Builde Manage and P oposalBuilde s, which
c ea e he accompanying p oposal.
Mo e speci ically, he Builde Manage selec s he mos app op ia e P oposal-
Builde o c ea e a new p oposal. Each P oposalBuilde implemen s a decision-
making algo i hm ha c ea es he p oposals o be sen o he o he pa y [18,
19, 33, 37]. The e o e, an au oma ed nego ia ion sys em may ha e se e al
P oposalBuilde s ha implemen di e en decision-making algo i hms (REQ
2.3), and he Builde Manage can choose dynamically which one should be
used o c ea e he p oposal (REQ 4.2), o ins ance, using a echnique like he
one desc ibed by Ros and Sie a [46]. The Builde Manage can also con igu e
he P oposalBuilde acco ding o he a o emen ioned nego ia ion policies.
Nego ia ion Con ex Da a
En i onmen al esou ce nego ia ionCon ex da a is ex ended by esou ce Ba -
gainingCon ex da a o s o e and p o ide in o ma ion abou he s a us o cu en
ba gaining nego ia ions.
Da a Model
The main con ibu ion o he NegoFAST da a model is ha i de ines a gene ic
model ha speci ies he main concep s o an au oma ed nego ia ion sys em
(p e e ences, ag eemen s, p oposals, nego ia ion messages, and in o ma ion
abou pa ies) independen ly o he o malism used o exp ess hem (see Fig-
u e 6). The concep s o he gene ic model a e pa ame ic. The e o e, c ea ing
a conc e e model in ol es binding he pa ame e s o he gene ic model o he
conc e e o malism used in he conc e e model. Consequen ly, p e e ences
and ag eemen s can be as exp essi e as necessa y, since any o malism (u ili y
unc ions, ules, name- alue pai s, cons ain s, uzzy cons ain s, combina ions
Figu e 6. P e e ence, Ag eemen , and P oposal Models

o a ibu es) can be used p o ided ha i complies wi h he ollowing condi-
ions (REQ 1.2):
• The o malismmus ex end heco espondinggene icelemen s:
IS a emen o o malisms used o exp ess p e e ences and in o ma ion
abou pa ies, and ITe m o o malisms used o exp ess ag eemen s o
p oposals (see Figu e 6).
• Fo eachpai o p e e encesmodelandag eemen model,anas-
sessmen mechanism (IAssessmen Mechanism) mus be de ined ha
e alua es and compa es wo p oposals ha ollow he gi en ag ee-
men model in he con ex o some p e e ences ha ollow he gi en
p e e ences model.
The gene ic model i sel can be ex ended o suppo ad anced ea u es. Fo
ins ance, e ms can be ex ended o add compensa ion clauses o p oposals can
be ex ended o include addi ional nego ia ion da a abou he e ms speci ied
in he p oposal.
The main concep s o he gene ic da a model a e he ollowing:
• P e e ences(in e aceIP e e encesDocumen ). These a e composed
o h ee se s o s a emen s (in e ace IS a emen ) abou ag eemen -
ela ed ea u es o he se ice o be p o ided, he equi emen s on
o he pa ies, and he nego ia ion guidelines ha he au oma ed
nego ia ion sys em mus ollow. To allow di e en o malisms o ex-
p ess p e e ences, hey a e pa ame e ized by he ype o s a emen s,
such as u ili y unc ions, cons ain s, pai name- alue, o ules
(REQ 2.5). No e ha he p e e ences abou he nego ia ion p ocess
( he nego ia ion guidelines) a e conside ed a he same le el as he
p e e ences abou he se ice o abou he o he pa ies. This enables
he de ini ion o p e e ences ha guide he beha io o he au oma -
ed nego ia ion sys em, such as deadline, numbe o ag eemen s o
each, and eage ness o each an ag eemen (REQ 2.6).
• Ag eemen sandp oposals(in e acesIAg eemen and IP oposal).
These a e composed o a se o e ms (REQ 1.1) and pa ame e ized
by he ype o e ms hey con ain. Te ms (in e ace ITe m) speci y
cons ain s o e some ag eemen - ela ed ea u es wi h which a
pa y mus comply and a e pa ame e ized by he ype o cons ain
hey enclose— o example, equali y (see he conc e e da a model in
Figu e 6), cons ain s o e one a ibu e, cons ain s o e se e al a -
ibu es, o uzzy cons ain s (REQ 2.4).
• Nego ia ionmessages(in e aceINego ia ionMessage). These
consis o a pe o ma i e— o example, p opose, accep , o commi
(in e ace pe o ma i e)—which mus be e ined by p o ocol-speci ic
ex ensions (in e ace Ba gainingpe o ma i e), and he con en s o he
message i sel (in e ace IMessageCon en ), which is a ag in e ace ha
indica es which elemen s may be pa o a nego ia ion message. Ne-
go ia ion messages a e pa ame e ized by he ype o hei con en s.
• In o ma ionabou pa ies(in e aceIPa yIn o ma ion). This mod-
els he public in o ma ion o e ed by he pa ies abou hei p e e -
ences and is ob ained by means o ole Inqui e . Like p e e ences, he
pa y in o ma ion is composed o wo di e en se s o s a emen s—
equi emen s and ea u es—and, also like p e e ences, i is pa am-
e e ized by he ype o s a emen used o exp ess hem. Fo ins ance,
in Figu e 6, a mix o weigh ed u ili y and name- alue pai s is used.
Valida ion
To p o e he soundness and use ulness o he NegoFAST e e ence a chi ec u e,
he NegoFAST amewo k was designed and implemen ed o help alida e
he con ibu ions by means o se e al case s udies, namely:
• Compu ingjobou sou cing(Case1). This case s udy is de eloped
in he con ex o a company ha ou sou ces compu ing powe o un
compu ing jobs (e.g., business in elligence jobs). I s goal is o imple-
men an au oma ed nego ia ion sys em o a compu ing job submi -
e ha nego ia es simul aneously wi h se e al job-hos ing se ices
o each an ag eemen on he esou ces and cos equi ed o execu e
one job.
• Compu ingjob-hos ingse ice(Case2). This case s udy is simila
o he p e ious one, bu i ocuses on he job-hos ing se ice ha
nego ia es wi h se e al compu ing job submi e s.
• E olu i eequilib ium(Case3). The goal o his case s udy is o
show how o in eg a e he NegoFAST amewo k wi h a Ja a ame-
wo k o gene ic algo i hms in o de o apply a known e olu i e
app oach o calcula e he equilib ium among s a egies.
• Schedulingmee ings(Case4). This case s udy ocuses on he imple-
men a ion o a mechanism o schedule mee ings by means o he
mul iagen nego ia ions desc ibed by Waine , Fe ei a, and Cons an-
ino [57].
The goal o hese case s udies is h ee old. Fi s , implemen ing he case
s udies shows ha he NegoFAST e e ence a chi ec u e can be ansla ed in o
an implemen a ion amewo k ha can be used o e ec i ely build au oma ed
nego ia ion sys ems.
Second, hese case s udies make i possible o check ha bo h he e e -
ence a chi ec u e and he amewo k suppo he equi emen s desc ibed in
he discussion o he backg ound o au oma ed nego ia ion sys ems. Table 3
summa izes he equi emen s co e ed by he di e en case s udies. No e ha
Requi emen s 2.4 and 2.5 a e co e ed by Case 1 and Case 4 oge he because
each o hem implemen s a di e en ag eemen and p e e ences model. (Al-
hough Requi emen s 3.2 and 4.5 ha e no been included in any o he case
s udies, he amewo k p o ides some speci ic ma ke model algo i hms [49].)
Finally, no alida ion has been done ega ding decommi men because he
cu en e sion p o ides e y li le suppo o i . This opic is no su icien ly
co e ed in he li e a u e and equi es u he esea ch be o e a comple e
amewo k can be de eloped.
Thi d, he ca e ul selec ion o hese ou case s udies makes i possible o
check desi able non unc ional p ope ies o he amewo k: Case 2 (compu ing
job-hos ing se ice) was chosen o es he eusabili y o he amewo k because
i is a simila scena io wi h di e ences ega ding he decision-making oles;
Case 3 (e olu i e equilib ium) was chosen o es he abili y o he amewo k
o be in eg a ed wi h o he sys ems, an impo an ea u e in a ealis ic scena io;
and Case 3 (e olu i e equilib ium) and Case 4 (scheduling mee ings) we e
chosen o es he adap abili y o he amewo k o new scena ios.
These case s udies we e implemen ed in wo s ages. Fi s , Augus , a p oo -
o -concep implemen a ion o he NegoFAST amewo k, was de eloped using
Ja a 1.5. Augus p o ides a e e ence implemen a ion o he in e aces and
gene ic da a model speci ied in he NegoFAST amewo k. Augus was used
o implemen au oma ed nego ia ion sys ems o he case s udies.
Compu ing Job Ou sou cing (Case 1)
This case s udy ocuses on he nego ia ion o an ag eemen be ween a com-
pu ing job submi e and se e al job-hos ing se ices ha need o ag ee on
he job o be execu ed, he esou ces equi ed, o scheduling equi emen s [2].
Figu e 7 illus a es his scena io, and Figu e 8 depic s i s componen diag am.
Fi s , he job submi e sends i s p e e ences, which include bo h equi emen s
abou he job execu ion and guidelines ega ding he nego ia ion p ocess, o i s
Table 3. Requi emen s Co e ed by Case S udies.
Requi emen Case 1 Case 2 Case 3 Case 4
(1.1) Suppo mul i e m nego ia ion p o ocols   
(1.2) Manage exp essi e p e e ences models   
(2.1) Suppo mul iple p o ocols  
(2.2) Nego ia e nego ia ion p o ocol  
(2.3) Suppo mul iple nego ia ion in elligen   
algo i hms
(2.4) Suppo mul iple ag eemen models  
(2.5) Suppo mul iple p e e ences models  
(2.6) Allow use p e e ences abou nego ia ion   
p ocess
(3.1) Manage di e en ypes o knowledge 
abou pa ies
(3.2) Ga he in o ma ion om di e en sou ces
(3.3) Build analysis‑based models 
(4.1) Suppo se e al simul aneous nego ia ions  
(4.2) Selec in elligence algo i hms dynamically 
(4.3) Suppo decommi men
(4.4) Supe ised c ea ion o ag eemen s 
(4.5) Build ma ke models
au oma ed nego ia ion sys em. In his case s udy, ou guidelines a e de ined
o con ol he nego ia ion p ocess: nego ia ion deadline, numbe o ag eemen s
o each, eage ness o each an ag eemen , and minimum u ili y h eshold.
Second, when he au oma ed nego ia ion sys em ecei es e e ences o job-
hos ing se ices, i s a s bila e al nego ia ions wi h hem. When an ag eemen
is eached, he au oma ed nego ia ion sys em sends he ag eemen o he job
Figu e 7. Compu ing Job Submission Scena io
Figu e 8. Au oma ed Nego ia ion Sys em o Compu ing Job
Submission
and he use o a P o ocolnego ia o (REQ 2.2). Ce ain o he ea u es a e less
han common in cu en s a e-o - he-a nego ia ion amewo ks, such as he
abili y o exp ess use p e e ences abou he nego ia ion p ocess (REQ 2.6),
suppo o se e al simul aneous nego ia ions (REQ 4.1), and he dynamic
selec ion o decision-making algo i hms (REQ 4.2). The only equi emen ha
is no ully suppo ed by NegoFAST is decommi men . In he cu en e sion
o NegoFAST, he decommi men suppo is nai e. The main eason is ha
decommi men is s ill a no el opic ha dese es u he esea ch be o e being
in eg a ed in o a amewo k.
On he basis o he NegoFAST e e ence a chi ec u e, a so wa e amewo k
has been designed and implemen ed, and i was used in his s udy o imple-
men au oma ed nego ia ion sys ems o se e al case s udies ha co e mos
o he a o emen ioned equi emen s. These case s udies ha e also made i
possible o check desi able non unc ional p ope ies o he amewo k, such
as eusabili y, in eg abili y, and adap abili y.
The con ibu ions made by his a icle will help p ac i ione s since hose
con ibu ions se he equi emen s and ounda ions o implemen au oma ed
nego ia ion sys ems. They also will enable esea che s o ocus on a pa icula
pa o an au oma ed nego ia ion sys em (e.g., wo ld modeling algo i hms
o esponse gene a ion algo i hms), while he e e ence a chi ec u e b idges
he gap be ween he cu en au oma ed nego ia ion echniques (p o ocols,
algo i hms, models), and, hence, eases he euse o cu en esul s in his ield.
The implemen a ion o he so wa e amewo k p o ides a ha ness in which
au oma ed nego ia ion echniques can be es ed om an empi ical poin o
iew. Fo ins ance, esea che s who wo k on decision-making algo i hms o
c ea e ag eemen p oposals o schedule mee ings can p oceed as ollows: Fi s ,
hey can es hei algo i hms by in eg a ing hem as new P oposalBuilde s in
he sys em de ailed in he discussion o scheduling mee ings, wi hou ha ing
o eimplemen he whole au oma ed nego ia ion sys em again. Second, hey
can compa e hei algo i hms wi h o he s ha ha e al eady been in eg a ed
in o he amewo k (e.g., he laconic, ego is ic, and decei ing algo i hms).
Thi d, hey can analyze how well he algo i hms pe o m in coo dina ion wi h
se e al di e en complemen a y models (e.g., how well hey pe o m using
di e en wo ld-modeling algo i hms).
ReFeReNCeS
1. Ab an, A.; Moo e, J.W.; Bou que, P.; and Dupuis, R. Guide o he so wa e
Enginee ing Body o Knowledge. Los Alami os, CA: IEEE Compu e Socie y
P ess, 2004.
2. And ieux, A.; Czajkowski, K.; Dan, A.; Keahey, K.; Ludwig, H.; Naka a,
T.; P uyne, J.; Ro ano, J.; Tuecke, S.; and Xu, M. WS-Ag eemen ecommen-
da ion. Ma ch 2007, www.g id o um.o g/documen s/GFD.107.pd .
3. Ash i, R.; Rahwan, I.; and Luck, M. A chi ec u es o nego ia ing agen s.
In V. Ma ík, J.P. Mülle , and M. Pechoucek, M. (eds.), Mul i-agen sys ems
and applica ions III: 3 d In e na ional Cen al and Eas e n Eu opean Con e ence
on Mul i-agen sys ems. Heidelbe g: Sp inge , 2003, pp. 136–146.

4. Ba dhan, I.R.; Whi ake , J.; and Mi has, S. An eceden s o business
p ocess ou sou cing in manu ac u ing plan s. In R.H. Sp ague (ed.), 39 h
annual hawaii In e na ional Con e ence on sys em sciences. Los Alami os, CA:
IEEE Compu e Socie y, 2006, pp. 68–69.
5. Ba dhan, I.; Demi kan, H.; Kannan, P.K.; Kau man, R.J.; and
Sougs ad, R. An in e disciplina y pe spec i e on IT se ices managemen
and se ices science. Jou nal o Managemen In o ma ion sys ems, 26, 4 (sp ing
2010), 13–65.
6. Ba olini, C.; P eis , C.; and Jennings, N.R. A so wa e amewo k o
au oma ed nego ia ion. In R. Cho en, A. Ga cía, C. Lucena, and A. Ramon-
o sky (eds.), so wa e Enginee ing o Mul i-agen sys ems III. Be lin: Sp ing-
e Ve lag, 2005, pp. 213–235.
7. Benyouce , M., and Ve ons, M.-H. Con igu able e-nego ia ion sys ems
o la ge scale and anspa en decision making. G oup decision and nego ia-
ion, 17, 3 (May 2008), 211–224.
8. Bichle , M. An expe imen al analysis o mul i-a ibu e auc ions. decision
suppo sys ems, 29, 3 (Oc obe 2000), 249–268.
9. Bleis ein, S.J.; Cox, K.; Ve ne , J.M.; and Phalp, K. Requi emen s engi-
nee ing o e-business ad an age. Requi emen s Enginee ing Jou nal, 11, 1
(Decembe 2005), 4–16.
10. Ch is ensen, C.M., and Rayno , M.E. how o a oid Commodi iza ion. Bos-
on: Ha a d Business School P ess, 2003.
11. Coehoo n, R.M., and Jennings, N.R. Lea ning on opponen s p e e ences
o make e ec i e mul i-issue nego ia ion ade-o s. In M. Janssen, H.G. Sol,
and R.W. Wagenaa (eds.), six h In e na ional Con e ence on Elec onic Com-
me ce. New Yo k: ACM P ess, Oc obe 2004, pp. 59–68.
12. Dai, Q., and Kau man, R.J. Business models o In e ne -based B2B
elec onic ma ke s. In e na ional Jou nal o Elec onic Comme ce, 6, 4 (summe
2002), 41–73.
13. Dan, A.; Da is, D.; Kea ney, R.; Kelle , A.; King, R.P.; Kueble , D.; Lud-
wig, H.; Polan, M.; Sp ei ze , M.; and Yousse , A. Web se ices on demand:
WSLA-d i en au oma ed managemen . IBM sys ems Jou nal, 43, 1 (Janua y
2004), 136–158.
14. Demi kan, H.; Kau man, R.J.; Vayghan, J.A.; Fill, H.-G.; Ka agiannis, D.;
and Maglio, P.P. Se ice-o ien ed echnology and managemen . Elec onic
Comme ce Resea ch and applica ions, 7, 4 (win e 2008), 356–376.
15. Doba dzie , A. Ou come-based p icing: Incen i ising inno a ion-led
IT se ices. O um Knowledge Cen e , June 2008, h p://s o e.o um.com/
P oduc .asp?pid=38842&e =in aus.
16. Dujmo ic, J.J. A me hod o e alua ion and selec ion o complex ha d-
wa e and so wa e sys ems. In 22nd In e na ional Con e ence o he Resou ce
Managemen and Pe o mance E alua ion o En e p ise Cs. Tu ne s ille, NJ:
Compu e Measu emen G oup, 1996, pp. 368–378.
17. El a a y, A., and Layzell, P.J. A nego ia ion desc ip ion language. so -
wa e, P ac ice and Expe ience, 35, 4 (Ap il 2005), 323–343.
18. Fa a in, P.; Sie a, C.; and Jennings, N.R. Nego ia ion decision unc ions
o au onomous agen s. In e na ional Jou nal o Robo ics and au onomous sys-
ems, 24, 3–4 (1998), 159–182.
19. Fa a in, P.; Sie a, C.; and Jennings, N.R. Using simila i y c i e ia o
make ade-o s in au oma ed nego ia ions. a i icial In elligence, 142, 2 (De-
cembe 2002), 205–237.
20. Fa ima, S.S.; Woold idge, M.; and Jennings, N.R. An agenda-based
amewo k o mul i-issue nego ia ion. a i icial In elligence, 152, 1 (Janua y
2004), 1–45.
21. Fa ima, S.S.; Woold idge, M.; and Jennings, N.R. A compa a i e s udy o
game heo e ic and e olu iona y models o ba gaining o so wa e agen s.
a i icial In elligence Re iew, 23, 2 (Ap il 2005), 187–205.
22. F iedman, T. he Wo ld Is Fla : a B ie his o y o he wen y- i s Cen u y.
New Yo k: Fa a , S aus, & Gi oux, 2005.
23. F ølund, S., and Kois inen, J. Quali y-o -se ice speci ica ion in dis ib-
u ed objec sys ems. dis ibu ed sys ems Enginee ing, 5, 4 (Decembe 1998),
179–202.
24. Gimpel, H.; Ludwig, H.; Dan, A.; and Kea ney, B. PANDA: Speci ying
policies o au oma ed nego ia ions o se ice con ac s. In M.E. O lowska,
S. Wee awa ana, M.P. Papazoglou, and J. Yang (eds.), Fi s In e na ional Con-
e ence on se ice-o ien ed Compu ing (ICsoC 2003). Be lin: Sp inge -Ve lag,
2003, pp. 287–302.
25. Go e na o i, G. Rep esen ing business con ac s in RuleML. In e na ional
Jou nal Coope a i e In o ma ion sys ems, 14, 2–3 (June 2005), 181–216.
26. G e en, P.W.P.J.; Ludwig, H.; Dan, A.; and Angelo , S. An analysis o
Web se ices suppo o dynamic business p ocess ou sou cing. In o ma ion
& so wa e echnology, 48, 11 (No embe 2006), 1115–1134.
27. G egg, D.G., and Walczak, S. Auc ion ad iso : An agen -based online-
auc ion decision suppo sys em. decision suppo sys ems, 41, 2 (Janua y
2006), 449–471.
28. He, M.; Jennings, N.R.; and Leung, H.-F. On agen -media ed elec onic
comme ce. IEEE ansac ions on Knowledge and da a Enginee ing, 15, 4 (July/
Augus 2003), 985–1003.
29. Jennings, N.R.; Fa a in, P.; Lomuscio, A.R.; Pa sons, S.; Woold idge, M.;
and Sie a, C. Au oma ed nego ia ion: P ospec s, me hods and challenges.
G oup decision and nego ia ion, 10, 2 (Ma ch 2001), 199–215.
30. Jonke , C.; Robu, V.; and T eu , J. An agen a chi ec u e o mul i-a i-
bu e nego ia ion using incomple e p e e ence in o ma ion. au onomous
agen s and Mul i-agen sys ems, 15, 2 (Oc obe 2007), 221–252.
31. Ka p, A.H. Rules o engagemen o au oma ed nego ia ion. In B. Bena-
allah and C. Goda (eds.), Fi s IEEE In e na ional Wo kshop on Elec onic
Con ac ing. Los Alami os, CA: IEEE Compu e Socie y, July 2004, pp. 32–39.
32. Kim, J.B., and Sege , A. A Web se ices-enabled ma ke place a chi ec-
u e o nego ia ion p ocess managemen . decision suppo sys ems, 40, 1
(July 2005), 71–87.
33. Kowalczyk, R. Fuzzy e-nego ia ion agen s. so Compu ing, 6, 5 (Augus
2002), 337–347.
34. Li, C.; Giampapa, J.; and Syca a, K. Bila e al nego ia ion decisions wi h
unce ain dynamic ou side op ions. IEEE ansac ions on sys ems, Man, and
Cybe ne ics, Pa C: applica ions and Re iews, 36, 1 (2006), 31–44.
35. Ludwig, A.; B aun, P.; Kowalczyk, R.; and F anczyk, B. A amewo k
o au oma ed nego ia ion o se ice le el ag eemen s in se ices g ids.
In C. Bussle and A. Halle (eds.), Business P ocess Managemen Wo kshops.
Be lin: Sp inge Ve lag, 2005, pp. 89–101.
36. Ludwig, H.; Dan, A.; and Kea ney, R. C emona: An a chi ec u e and
lib a y o c ea ion and moni o ing o WS-Ag eemen . In M. Aiello, M.
Aoyama, F. Cu be a, and M.P. Papazoglou (eds.), second In e na ional Con-
e ence on se ice-o ien ed Compu ing (ICsoC 2004). New Yo k: ACM P ess,
No embe 2004, pp. 65–74.
37. Luo, X.; Jennings, N.R.; Shadbol , N.; Leung, H.-F.; and Lee, J.H. A uzzy
cons ain based model o bila e al, mul i-issue nego ia ions in semi-com-
pe i i e en i onmen s. a i icial In elligence, 148, 1–2 (Augus 2003), 53–102.
38. Nash, J.F. The ba gaining p oblem. Econome ica, 18, 2 (Ap il 1950),
155–162.
39. Nguyen, T.D., and Jennings, N.R. Managing commi men s in mul iple
concu en nego ia ions. Elec onic Comme ce Resea ch and applica ions, 4, 4
(win e 2005), 362–376.
40. O danini, A. Wha d i es ma ke ansac ions in B2B exchanges? Com-
munica ions o he aCM, 49, 4 (Ap il 2006), 89–93.
41. Papazoglou, M.P. The challenges o se ice e olu ion. In Z. Bellahsene
and M. Léona d (eds.), 20 h In e na ional Con e ence on ad anced In o ma ion
sys ems Enginee ing (CaisE 2008). Be lin: Sp inge , June 2008, pp. 1–15.
42. Papazoglou, M.P., and an den Heu el, W.-J. Se ice o ien ed a chi ec-
u es: App oaches, echnologies and esea ch issues. In e na ional Jou nal on
Ve y la ge da a Bases, 16, 3 (July 2007), 389–415.
43. Pau obally, S.; Tamma, V.; and Woold idge, M. A amewo k o Web
se ice nego ia ion. aCM ansac ions on au onomous and adap i e sys ems,
2, 4 (No embe 2007), 14.
44. Plumme , D.C.; Smulde s, C.; Fie ing, L.; Na is, Y.V.; Mingay, S.; D i e ,
M.; Fenn, J.; McLellan, L.; and Wilson, D. Ga ne ’s op p edic ions o IT
o ganiza ions and use s, 2008 and beyond. S am o d, CT: Ga ne , 2008,
www.ga ne .com/i /page.jsp?id=593207.
45. Rinde le, S., and Benyouce , M. Towa ds he au oma ion o e-nego ia ion
p ocesses based on Web se ices. In A.H.H. Ngu, M. Ki su egawa, E.J.
Neuhold, J.-Y. Chung, and Q.Z. Sheng (eds.), 6 h In e na ional Con e ence
on Web In o ma ion sys ems Enginee ing. Be lin: Sp inge Ve lag, 2005, pp.
443–453.
46. Ros, R., and Sie a, C. A nego ia ion me a s a egy combining ade-o
and concession mo es. au onomous agen s and Mul i-agen sys ems, 12, 2
(Ma ch 2006), 163–181.
47. Rus , R.T., and Kannan, P.K. E-se ice: A new pa adigm o business in
he elec onic en i onmen . Communica ions o he aCM, 46, 6 (June 2003),
36–42.
48. Sandholm, T.W., and Lesse , V.R. Le eled commi men con ac s and
s a egic b each. Games and Economic Beha io , 35, 1–2 (Ap il 2001), 212–270.
49. Sim, K.M., and Choi, C.Y. Agen s ha eac o changing ma ke si ua-
ions. IEEE ansac ions on sys ems, Man and Cybe ne ics, Pa B, 33, 2 (Ap il
2003), 188–201.
50. Sim, K.M., and Wang, S.Y. Flexible nego ia ion agen wi h elaxed deci-
sion ules. IEEE ansac ions on sys ems, Man and Cybe ne ics, Pa B, 34, 3
(June 2004), 1602–1608.
51. Smi h, J.M., and P ice, G.R. The logic o animal con lic . na u e, 246 (No-
embe 1973), 15–18.
52. S öbel, M. Design o oles and p o ocols o elec onic nego ia ions.
Elec onic Comme ce Resea ch, 1, 3 (July 2001), 335–353.
53. Su, S.Y.W.; Huang, C.; Hamme , J.; Huang, Y.; Li, H.; Wang, L.; Liu, Y.;
Pluempi iwi iyawej, C.; Lee, M.; and Lam, H. An In e ne -based nego ia ion
se e o e-comme ce. In e na ional Jou nal on Ve y la ge da a Bases, 10, 1
(Augus 2001), 72–90.
54. Tu, M. T.; Seebode, C.; G i el, F.; and Lame sdo , W. DynamiCS: An
ac o -based amewo k o nego ia ing mobile agen s. Elec onic Comme ce
Resea ch, 1, 1–2 (Feb ua y 2001), 101–117.
55. Uma , A. IT In as uc u e o enable nex gene a ion en e p ises. In o ma-
ion sys ems F on ie s, 7, 3 (July 2005), 217–256.
56. Va go, S.L., and Lusch, R.F. E ol ing o a new dominan logic o ma -
ke ing. Jou nal o Ma ke ing, 68, 1 (Janua y 2004), 1–17.
57. Waine , J.; Fe ei a, P.R.; and Cons an ino, E.R. Scheduling mee ings
h ough mul i-agen nego ia ions. decision suppo sys ems, 44, 1 (No em-
be 2007), 285–297.
58. Wall, Q. Re hinking SOA go e nance. O acle Inc., 2008, h p://quin on
wall.com/wp-con en /uploads/2008/08/ e hinking-soa-go e nance.pd .
59. Zeng, D., and Syca a, K. Bayesian lea ning in nego ia ion. In e na ional
Jou nal human-Compu e s udies, 48, 1 (Janua y 1998), 125–141.
60. Zhu, L.; Leach, P.; Jagana han, K.; and Inge soll, W. The simple and
p o ec ed gene ic secu i y se ice applica ion p og am in e ace (GSS-API)
nego ia ion mechanism (RFC 4178). Ne wo k Wo king G oup, Oc obe
2005, h p:// ools.ie .o g/h ml/ c4178.
MANUEL RESINAS ([email p o ec ed]) is a lec u e on so wa e enginee ing a he Uni-
e si y o Se ille, Spain, whe e he ecei ed his Ph.D. in 2008. His esea ch in e es s
include au oma ed nego ia ion and au oma ed analysis o se ice ag eemen s and hei
ela ionship wi h business p ocesses.
PABLO FERNÁNDEZ ([email p o ec ed]) is a lec u e on so wa e enginee ing a he
Uni e si y o Se ille, whe e he is wo king owa d his Ph.D. His esea ch in e es s
ocus on au oma ed ading.
RAFAEL CORCHUELO ([email p o ec ed]) is a eade in so wa e enginee ing in he
Depa men o Compu e Languages and Sys ems o he Uni e si y o Se ille, whe e
he ecei ed his Ph.D. He has been he leade o he uni e si y’s Resea ch G oup on
Dis ibu ed Sys ems since 1997. His esea ch in e es s ocus on he in eg a ion o Web
da a islands; p e iously, he wo ked on mul ipa y in e ac ion and ai ness issues.