ANDALUCÍA ASSESSES THE INVESTMENT NEEDED TO DEPLOY A
FIBER-OPTIC NETWORK
Pablo Co es
[email p o ec ed]
Ingenie ía de O ganización. Escuela Supe io de
Ingenie os. Se ille Uni e si y. Camino de los
Descub imien os s/n, 41092. Se ille (Spain)
Jesus Muñuzu i
[email p o ec ed]
Asociación de In es igación y Coope ación Indus ial
de Andalucía (AICIA). Camino de los
Descub imien os s/n, 41092. Se ille (Spain)
Luis Onie a
[email p o ec ed]
Asociación de In es igación y Coope ación Indus ial
de Andalucía (AICIA). Camino de los
Descub imien os s/n, 41092. Se ille (Spain)
Juan La añe a
[email p o ec ed]
Ingenie ía de O ganización. Escuela Supe io de
Ingenie os. Se ille Uni e si y. Camino de los
Descub imien os s/n, 41092. Se ille (Spain)
Juan M. Vozmediano
[email p o ec ed]
Ingenie ía Telemá ica Escuela Supe io de Ingenie os.
Se ille Uni e si y. Camino de los Descub imien os s/n,
41092. Se ille (Spain)
Jose C. Ala con
[email p o ec ed]
Sec e a ía Gene al de Sis emas de In o mación y
Telecomunicaciones.
Conseje ía de la P esidencia. Jun a de Andalucía.
1
Abs ac .
The se up o ibe -op ic elecommunica ion ne wo ks in ol es high in es men e o s.
The Regional Go e nmen o Andalusia assigned us he de elopmen o a ool capable
o e alua ing he deploymen cos o a ne wo k ha was no o be limi ed only o
connec ing la ge ci ies, bu also o include smalle owns, in o de o p e en hem om
s aying behind he p og ess o he In o ma ion Socie y. The Andalusian egional
Go e nmen aimed o deploy a ne wo k capable o accessing mos o he municipali ies
in he egion, e en hose municipali ies ha could no be p o i able om a mone a y
pe spec i e. We de eloped a nonlinea ma hema ical p og amming model wi h special
ocus on he in es men cos s. The cos s included he pa s co esponding o he ci il-
enginee ing wo ks, as well as hose ela ed o he elema ic link deploymen . The
solu ion o such a complex p oblem was ound by a gene ic algo i hm, which was
p e iously es ed wi h a se o ial p oblems. The esul s we e used o pe suade p i a e
companies o expand hei ibe -op ic ne wo ks o each small owns.
The concep o in o ma ion highway has come o mean high-speed elecommunica ion
ne wo ks. These ne wo ks equi e an in as uc u e ha will suppo he long-dis ance
exchange o in o ma ion ia elec onic means, equi ed o di e en business ac i i ies.
The u u e in o ma ion supe -highway will ha e o combine in a single in as uc u e
he bes ea u es o all he equi ed ne wo ks; i will need a bandwid h ha will
gua an ee he quali y necessa y o ele ision, in e ac ion among cus ome s, and secu e
inancial and comme cial ansac ions.
The de elopmen o swi ching echniques, bandwid h sha ing, and ibe -op ic
ansmission makes possible ne wo ks ha p o ide high quali y, and in eg a e many
se ices unde he same use in e ace.
2
To es ablish a s anda d o he in o ma ion highway, in he 1980s, he In e na ional
Telecommunica ion Union (ITU) p oposed deploying he ibe -op ic ne wo k wi h
ansmission by means o synch onous digi al hie a chy (SDH), and swi ching ia he
asynch onous ans e mode (ATM) (Sex on and Reid, 1997).
While Eu ope adop ed he SDH s anda d, he USA adap ed he synch onous op ical
ne wo k (SONET), p oposed by he Ame ican Na ional S anda ds Ins i u e (ANSI),
which is a simila bu somewha di e en s anda d ( able 1). The ollowing able depic s
he ela ion be ween s anda ds, as well as he co esponding equi alen op ical le el and
he capaci y in Mbps.
Op ical le el Capaci y (Mbps) SDH (ITU) a SONET (ANSI) b
OCc-1 51.84 STSd-1 -
OC-3 155.52 STS-3 STMe-1
OC-9 466.56 STS-9 -
OC-12 622.08 STS-12 STM-4
OC-18 933.12 STS-18 -
OC-24 1244.16 STS-24 -
OC-36 1866.24 STS-36 -
OC-48 2488.32 STS-48 STM-16
Table 1: The Eu opean Union and he USA ha e adop ed di e en
ansmission s anda ds o ibe -op ic ne wo ks. Depic ed he e is he ela ion
be ween bo h s anda ds, as well as he co esponding equi alen op ical le el
and he capaci y in Mbps (a Synch onous Digi al Hie a chy - In e na ional
Telecommunica ion Union; b Synch onous Op ical NETwo k - Ame ican
Na ional S anda ds Ins i u e; c Op ical Ca ie le el; d Synch onous
T anspo Signal le el; e Synch onous T anspo Module le el).
3
In Spain, he deploymen o cable ne wo ks is a less ex ensi e han i is in o he
wes e n coun ies such as he US and Belgium. The Andalusian Regional Go e nmen
has been p omo ing he de elopmen o ibe -op ic ne wo ks.
The Gene al Di ec o a e o he In o ma ion Socie y in he Jun a de Andalucía wan ed a
ool o e alua ing he cos s o se ing up a ibe -op ic ne wo k ac oss he Andalusia
egion. I wan ed ha ne wo k o connec bo h la ge ci ies and small owns so ha hey
would no all behind in he in o ma ion socie y. The Associa ion o Indus ial
Resea ch and Coope a ion in Andalusia (AICIA), oge he wi h he Depa men o
Managemen Enginee ing o he Uni e si y o Se ille wo ked oge he o de elop a ool
o es ima ing in es men equi emen s. The Jun a de Andalucía wan ed o es ablish a
ne wo k capable o accessing mos o he municipali ies in he egion, e en hose ha
could no be p o i able om a mone a y pe spec i e. I also wan ed o use he e alua ion
esul s and he ne wo k speci ica ions o pe suade p i a e companies o expand hei
own ne wo ks by using links in he “adminis a ion ne wo k” o each small owns.
Ou echnical s udy is he basis o he egional go e nmen ’s cu en pilo p ojec on
he deploymen o a uni e sal elecommunica ion ne wo k in he Andalusia Region. On
Ma ch 10 h, 2004, he gene al manage o he Gene al Di ec o a e o In o ma ion
Socie y o he Jun a de Andalucía desc ibed he p ojec o he p ess. Al hough i
o iginally wan ed o pe suade companies o make a majo e o o es ablish he
ne wo k, he egional go e nmen is now building he ne wo k wi h i s own unds and
wi h unds om he Eu opean Union. The egional adminis a ion is s ill ying o
pe suade a ious companies o ake pa in es ablishing he ne wo k.
G aph heo y and o he OR echniques ha e p o ed use ul in ne wo k deploymen (Cox
e al., 1993). We de eloped a model ha ep esen s a elecommunica ions ne wo k
4
easonably well, and a p ocess o deploying a cos -e ec i e ne wo k deploymen o a
geog aphic a ea he size o a egion. Resea che s who made impo an con ibu ions in
his a ea we e Balak ishnan e al. (1998), who conside ed he op imal capaci y
in es men p oblem, and Chang e al. (1995) and Saniee (1995), who s udied he
mul ipe iod case. Al hough hei wo k was impo an , i did no e lec he echnological
complexi y o deploying a elecommunica ions ne wo k, in he opinion o ou
echnological pa ne s om he Uni e si y’s Telema ic Depa men and om indus y
pa ne s. As Michael Pidd (1999) asked, wha is mos impo an : adap ing eali y o he
models o adap ing he models o eali y? To inco po a e ou echnological pa ne s’
commen s and a e discussing hem, we emodeled he adi ional mul icommodi y
low model, cons uc ing an exhaus i e and ex emely complex model (Appendix).
Ano he app oach o dealing wi h he p oblem is eal op ions heo y used in in es men
analysis. I is an especially in e es ing app oach because i conside s sunk in es men
cos s. Keppo (2003) uses his app oach o deal wi h a elecommunica ion capaci y
in es men p oblem, bu we did no ollow his line.
S a ing a mul idisciplina y p ojec : he discussion
The eam esponsible o he p ojec included enginee s in he elema ic enginee ing
ield, enginee s wo king in ope a ions esea ch, enginee s om he egional
go e nmen , p o essionals om he elecommunica ions indus y, and people om o he
academic backg ounds. Such a mul idisciplina y eam has g ea po en ial i i s membe s
can manage o each ag eemen s.
The main p oblem wi h eamwo k is ha he di e en eam membe s y o eplica e
me hods hey ha e ound success ul in he pas and in which hey specialize. Thus,
while he ope a ions esea ch specialis s ied o con ince he o he eam membe s o
5

he need o apply models based on mul icommodi y lows, he elema ic enginee s and
he elecommunica ion p o essionals o ed o eplica ing well-known ne wo k
s uc u es, eeling no need o complex op imiza ion me hods. E e y hing seemed o be
se in s one, and he only hing le was o look a he lis o p o ide s. Howe e , he
egional adminis a o s wan ed o minimize he o e all cos . Besides, i wan ed o a oid
a mul i-s age deploymen ; and i s own echnical s a was o implemen he p ojec .
A e he ini ial mee ings in which consensus seemed emo e, he membe s o he eam
s a ed o eally unde ake eamwo k. Those amilia wi h ope a ions esea ch began o
unde s and he pe o mance and echnological composi ion o ne wo ks, and hose in
elema ics unde s ood ha hey we e assuming he me hods hey knew we e easible
based on commonplace heu is ics, bu ha hose me hods migh p oduce a solu ion ha
was a om op imal, and could be o lowe quali y han o he un es ed solu ions. The
egional adminis a ion wan ed a low-cos ne wo k wi h ex ensi e co e age ha would
each he mos unde de eloped owns. I wan ed o c ea e an in o ma ion socie y, and
ob ain a poli ically co ec wel a e solu ion. The Eu opean Union imposed some o
hese equi emen s in p o iding Eu opean S uc u al Funds: he ne wo k had o each
e en hose owns whe e deploymen o a ibe -op ic ne wo k would be o li le o no
in e es o p i a e companies.
The eam membe s had o come o ag eemen on a se ies o ma e s. Fi s , i ag eed he
ne wo k s uc u e would be hie a chical, would include he exis ing elecommunica ion
ne wo ks, would ha e wo le els, and would co e la ge g oups o owns and inc ease
he le el o mul iplexing wi hin he e i o ial bounda ies o Andalusia. Second, he
eam membe s ag eed ha he cos model would be buil as a piecewise linea unc ion
( his objec i e was una ainable, acco ding o he elema ic enginee s). Thi d, he model
would ep esen he ne wo k s uc u e and objec i e unc ion, and he cons uc ion-
6
ela ed es ic ions ypical o his ype o ne wo k. The eam membe s also ag eed ha
he g ade o su i abili y (pa hs connec ing a pai o ac i e nodes) would be 2,
ollowing es ablished p ac ice in he indus y. Once he eam buil he model, i had o
ind a solu ion me hod. The OR membe s pe suaded he o he eam membe s o use
gene ic algo i hms.
Wi hou doub , he mos complex ask o any in he p ojec was eaching ag eemen o
build a model ha closely ep esen ed eali y, and adap ing o i , ins ead o ying o
adap eali y o he model.
The hie a chical s uc u e o he model
P i a e companies in he cable elecommunica ions sec o se up hei ne wo ks using
hie a chical s uc u es wi h di e en le els (Ma inez and Pa ames, 2001). Modeling
ne wo ks wi hou hie a chical le els is an exe cise in ma hema ical abs ac ion, bu no
a ealis ic applica ion in he elecommunica ions indus y. To ep esen hie a chical
le els, one mus in oduce addi ional complexi y in he models. Di e en g aphs a e
associa ed wi h he a ious hie a chical le els in e ms o nodes and links, and hey a e
in e ela ed wi h he o he g aphs.
In designing elecommunica ion ne wo ks, ope a ion esea che s in di e en heo e ical
a eas ha e sough o use well known p oblems, such as he hub-loca ion p oblem, he
cons ained minimum-spanning- ee p oblem, he S eine ee p oblem o he
capaci a ed-mul icommodi y- low p oblem o sol e he eal p oblem. We de eloped a
model ha was less compac and elegan , bu close o eali y.
We conside ed h ee hie a chical le els. The lowes was passi e nodes ep esen ing
simple in e sec ions in he ne wo k in as uc u e. No da a would be inse ed o
e ie ed in he ne wo k ia hese nodes. A he nex hie a chical le el we e owns, o
7
u ban nodes, used o inco po a ing and ecei ing da a om he ne wo k. Finally, a he
highes le el we e swi ching cen e s, o hubs (assigned o de ined geog aphical egions
wi hin he ne wo k deploymen a ea), con aining ansmission equipmen , such as he
digi al c oss-connec sys em (DCS).
Ou model con ains, hen, wo o e lapping ne wo k s uc u es: a physical ne wo k and a
logical ne wo k. The physical ne wo k de e mines he physical loca ion o di ches, ha
is, i ep esen s needed ci il-enginee ing wo k. I includes all he nodes in he h ee
hie a chical le els. The logical ne wo k, on he o he hand, is based on wo hie a chical
le els. The lowes logical le el ( he low-speed ne wo k) connec s he u ban nodes (in
he second hie a chical le el) and he hubs (in he hi d le el). I includes links ha
handle all he communica ions be ween he u ban node and i s egional swi ching cen e
(hub). The highes logical le el ( he high-speed ne wo k), handles communica ion
be ween egional swi ching cen e s (hubs). I includes sha ed links ha ca y
communica ions om di e en egions oge he , a swi ching cen e ha sepa a es hem
and e ou es hem o hei inal des ina ions (Figu e 1).
8
Figu e 1: Ne wo k wi h h ee egions. This ibe -op ic communica ions ne wo k
has a hie a chical s uc u e o nodes and links o low ansmission, ea u ing a
ing-s a logical s uc u e, equi alen o a ing- ee physical s uc u e. Fo
in a egional communica ion be ween wo u ban nodes T
1 and T’1 ,in o ma ion
lows om T1 o H1 ollow he dedica ed in a egional link om T1 o H1 included
in he logical ne wo k, which con ains he condui s connec ing T1 and 1, and 1 and
H1 in he physical ne wo k. The swi ching cen e H1 sends he in o ma ion o i s
co ec des ina ion T’1 , ia he link (H1 , T’1), which includes condui s (H1 , 1) and
(1 , T’1). Fo in e egional communica ion be ween u ban node T1 and u ban node
T2 , in o ma ion lows om T1 o H1 ollow he dedica ed in a egional link (T1 ,
H1). A e swi ching in hub H1, hey go h ough he in e egional sha ed link (H1 ,
H2), which can ca y se e al in e egional communica ions oge he a he same
ime, and which includes condui s (H1 , 2) and (2 , H2). Finally, a e swi ching in
H2 , he communica ion a els o he inal node T2 ia he dedica ed in a egional
link (H2 , T2), which includes condui s (H2 , 3), (3 , 4) and (4 , T2) o he physical
ne wo k.
The link connec ion (Figu e 1) allows any ne wo k en i y o communica e wi h any
o he . In p ac ice, ne wo k planne s commonly use second deg ee edundancy le els o
he uppe -le el logical ne wo k (c ea ing a ing s uc u e), connec ing u ban nodes (o a
lowe hie a chical le el) ia a s a - ype logical connec ion, which co esponds o a ee-
ype physical s uc u e. Ne wo k ope a o s, howe e , habi ually ejec as un easible
hose pa hs o ansmi ing communica ions ha equi e oo many swi ching p ocesses
in hubs. By se ing his c i e ion hey in end o minimize he delay in he pa h, because
he swi ching p ocess en ails big delays (Appendix).
9
Ma ching delay le els [Appendix equa ion (12)]
The model (Appendix) can be classi ied as a mixed-in ege linea -p og amming
p oblem. Because he p oblem has a combina o y s uc u e, i is di icul o sol e. Fo
example conside he ealis ic case o a geog aphical a ea di ided in o i e egions, wi h
20 municipali ies (which imply abou 28 possible in e egional communica ions and 30
egional ones) and a g aph wi h 600 links including an a e age o 200 ibe -op ic links
associa ed wi h each possible hub loca ion inside he egion and h ee al e na i e pa hs
o each in e egional communica ion. The model would equi e 1,644 bina y a iables,
209,974 con inuous a iables, and 206,739 es ic ions, excluding non-nega i i y
condi ions.
To deal wi h hese dimensions, we ackled he p oblem using a p ocedu e based on
hie a chical decomposi ion. In he i s phase, we concen a ed on loca ing hubs and
ex ending he physical ne wo k wi h eliabili y condi ions, since mos o he cos
co e ed hose ac i i ies (up o 85 pe cen ). No e ha a ne wo k su i al o ailu es
needs eliabili y o e se e al o hei links, especially in he unk subne wo k. In he
second phase we conside ed he associa ed p oblem o ou ing demand and assigning
capaci y.
As one o he mos impo an decisions we needed o loca e hubs; hei loca ions would
s ongly in luence bo h he physical and he logical ne wo ks ( igu e 3).
16

65
4
3
2
1
7
165
4
3
2
B
7
65
4
3
2
1
C
7
65
4
3
2
1
D
7
3
165
4
2
B1 165
4
2
B2 16
4
2
B
3
3
5
4
3
2
C1 6
5
3
2
C2
65
4
3
2
C3
65
3
2
D
1
7
5
4
3
2
D
2
7
65
4
3
2
D3
7
A
Figu e 3: Conside ing ne wo k A which co e s h ee egions, B, C, and D a e
possible hub loca ions. Each p oduces di e en physical and logical ne wo ks. In
depic ing ne wo ks among hubs; we ha e omi ed he es o he nodes ( o
example, we include node 1 when i is he hub o egion 1, in s uc u e B). A le el
2 o he hie a chy, he municipali ies will be associa ed wi h hubs acco ding o he
hub s uc u e selec ed. Thus he hub in op ion B could suppo he g oups o links
o ex ending he ibe -op ic ne wo k shown in B1, B2, and B3. C would esul in
C1, C2 o C3, and D in D1, D2 o D3. Addi ional al e na i es o hub s uc u es
(up o 12) exis o loca ing swi ching cen e s. A e he hub s uc u e is ixed, one
can de e mine he exac numbe o easible links.
Because each hub s uc u e can suppo a numbe o se s o easible links, de e mining
he loca ions o hubs is a key issue. Ou solu ion p ocedu e elies on his ac . We
o mula ed a gene ic algo i hm in which he ch omosome encoding o di e en
17
indi iduals ep esen ed he di e en op ions o hub loca ions. We e alua ed he i ness
o each indi idual in he popula ion using a wo-phase hie a chical p ocess.
Thus, o each p oposed hub s uc u e, he i s phase consis ed o de e mining a
connec ed ne wo k wi h a κk su i abili y g ade. We sough a su i abili y g ade κk =2
o he in e egional ne wo k and a su i abili y g ade κk =1 o he in a egional one.
We would hus achie e he con igu a ion associa ed wi h he basic condui
in as uc u e ob ained by using wo- ee echniques wi h su i abili y condi ions,
combining S eine ees (Monma and Shallc oss 1988 and Monma e al. 1990).
In he second phase, we decided on he logical ne wo k in which he op ical ibe links
we e se and i s capaci y was se . To sol e his p oblem we o mula ed he Kuhn-Tucke
op imali y condi ions associa ed wi h he educed p oblem, in which es ic ions (5), (7)
and (9) do no need o be conside ed, since hey a e implici ly o ced a e we encode
he indi idual’s ch omosome in he gene ic popula ion. Associa ed wi h he op imali y
condi ions, we de eloped a ou ing-demand and capaci y-assignmen algo i hm ha
gua an eed i app oxima ed he op imal solu ion in each i e a ion (Co es e al 2001).
The esul s
Ini ially, we wan ed o compa e ou esul s wi h he esul s om o he simila egions in
Spain o Eu ope. Howe e , we could no because he p i a e companies ope a ing in
his a ea conside ed such in o ma ion con iden ial. Fu he mo e, he Adminis a ion
(Regional, Na ional o Eu opean) had no p omo ed p e ious expe iences wi h such
p ojec s in he Spanish o Eu opean con ex .
Howe e , we compa ed he ou come o ou model and algo i hm wi h he esul s om
he well-known wo-s ep heu is ic, (Sex on and Reid 1997), which p i a e companies
use when deploying elecommunica ions ne wo ks.
18
In his way, we alida ed ou solu ion algo i hms wi h 600 ial p oblems. We gene a ed
hese ial p oblems andomly among di e se anges, bu wi hin he p oblem cons ain s.
To compa e ou gene ic algo i hm wi h his wo-s ep heu is ic, we conduc ed
expe imen s (Table 5).
Fac o code Fac o desc ip ion
L
owe le el (code “-“)
U
ppe le el (code “+”)
1. N
N
umbe o nodes
N
≤ 60
N
> 60
2. DA
A
cs densi y
D
A ≤ 1/2
D
A > ½
3. DR
R
egion densi y
D
R ≤ 1/3
D
R > 1/3
4. DT Te minals densi y
D
T ≤ 1/2
D
T > ½
5. D
D
emand scena io
B
ase demand scena io
H
igh demand scena io
6. S
R
equi ed su i abili y
L
e el 2 in in e
- egional ne wo k
M
o e han le el 2 in in e
- egional ne wo k
Table 5: We analyzed he ac o s e ec using he Expe imen al Design Theo y. We
conside ed six ac o s, ou opological ac o s and wo echnological ac o s
di ided in o wo le els (a wo le els codi ica ion). The Expe imen al Design Theo y
s a es ha such codi ica ion gua an ee ha he a iance analysis coincides wi h he
minimum squa e e o analysis. A e ha , we used Limdep .7.0 (econome ic
so wa e om NLogi ) o adjus he mode and o analyze he e ec s o hese
ac o s on he cos o es ablishing he ne wo k and on he su i abili y le el
a ained (SLA).
19
The numbe o nodes (N) includes he passi e nodes as well as he municipali ies
(ac i e nodes).
We calcula ed a c densi y (DA) as A/MAXARC, whe e A ep esen s he numbe o
a cs in he unde lying anspo ne wo k and MAXARC he maximum numbe o
a cs possible o he numbe o nodes in he g aph.
We calcula ed he densi y o he egions (DR) as R/T, whe e R ep esen s he
numbe o egions in he geog aphical a ea and T he numbe o u ban nodes.
We calcula ed he densi y o e minals (DT) as T/N.
We classi ied ou esul s by anges acco ding o Table 6. The p oblem ange is gi en by
he numbe o nodes, he e minals densi y and he egion densi y.
P oblem ange Gene ic algo i hm Two-s ep heu is ic Gene ic heu is ic
Nodes Te minal densi y
(DT)
Region densi y
(DR) Cos Time Cos Time
Cos
educ ion A c densi y
(DA) SLA SLA
[0.1-0.6] 100% 100%
[0.25-0.4] 12.91 4.7 18.11 5.1 29%
(0.6-1.0] 100% 100%
[0.1-0.6] 98% 97%
[0.1-0.55]
(0.4-0.6] 11.12 5.1 14.23 5.2 22% (0.6-1.0] 100% 100%
[0.1-0.6] 97.5% 95%
[0.25-0.4] 20.76 23.8 28.78 24.5 28% (0.6-1.0] 98.3% 98%
[0.1-0.6] 88% 86%
≤ 40
(0.55-1.0]
(0.4-0.6] 20.21 60.6 35.12 46.8 42% (0.6-1.0] 99% 98%
[0.1-0.6] 98.9% 98%
[0.25-0.4] 22.08 37.8 34.50 62.1 36% (0.6-1.0] 99% 99%
[0.1-0.6] 98.5% 92.9%
[0.1-0.55]
(0.4-0.6] 29.47 92.6 45.24 95.5 35% (0.6-1.0] 100% 98%
[0.1-0.6] 92.5% 86.8%
[0.25-0.4] 43.91 208.8 60.74 150.1 28% (0.6-1.0] 96% 92.8%
(40-80]
(0.55-1.0]
(0.4-0.6] 54.56 592.8 81.65 303.2 33% [0.1-0.6] 91% 90%
20
(0.6-1.0] 94.3% 89.5%
[0.1-0.6] 99.5% 97%
[0.25-0.4] 26.57 101.2 62.58 147 57.5% (0.6-1.0] 100% 100%
[0.1-0.6] 99.8% 96%
[0.1-0.55]
(0.4-0.6] 23.72 80 33.21 99 28.6% (0.6-1.0] 100% 100%
[0.1-0.6] 93.2% 90.4%
[0.25-0.4] 54.53 560.5 82.05 495.8 33.5% (0.6-1.0] 95.3% 91.1%
[0.1-0.6] 86.4% 80.3%
(80-100]
(0.55-1.0]
(0.4-0.6] 62.45 1720 126.42 952.3 50.6% (0.6-1.0] 87.5% 81.1%
Table 6: The esul s show ha he gene ic algo i hm a ains be e esul s han he
wo-s ep heu is ic o ne wo ks in all anges. Fo la ge ne wo ks, he gene ic
algo i hm educes cos s mo e han o small ne wo ks. Bo h e minal and egion
densi ies eme ged as impo an explaining ac o s. The densi y-o -a cs pa ame e
does no signi ican ly a ec cos o ime o execu ion; bu i is he main ac o
a ec ing SLA. Compu a ion ime o he wo-s ep heu is ic is lowe han ha o
he gene ic algo i hm speci ic anges because i is based on simple ules and does
no use i e a i e calculus as gene ic algo i hms do. Howe e , we ob ained all he
esul s wi hin easible bounded execu ion imes.
In able, cos s a e in million o eu os, a e age execu ion ime a e in seconds.
The su i abili y le el a ained pe cen age is he inal su i abili y le el a ained
di ided in o he su i abili y le el equi ed.
We ound ha co e age is associa ed wi h he densi y o he e minals, since co e age is
calcula ed as he a io be ween he ac i a ed nodes in he ne wo k ( e minals, T) and he
o al numbe o nodes (N) in he ne wo k. The egion’s densi y indica es he
mul iplexing le el in he ne wo k: o a high alue o DR, we will ha e a low le el o
mul iplexing, and o a low alue o DR, we will ha e a high le el o mul iplexing.
Based on ou esul s, we d ew some gene al conclusions ( igu e 4).
21

Co e age (nodes)
810 12 14 16 18 20 22
0
1000
2000
3000
4000
5000
10 12 14 16 18 20 22
0
0,5
1
1,5
2
Co e age inc ease
Cos inc ease
Numbe o egions
3 4 5 6 7 8 9 10
2500
3000
3500
4000
4500
Good-enginee ing Unde -enginee ing
O e -enginee ing Demand inc ease
Cos inc ease
Figu e 4: The ne wo k co e age is in a linea ela ionship wi h he cos o
deploying i , al hough he cos inc emen s a e smalle han he inc emen s in
ne wo k co e age. On he o he hand, he mul iplexing a ec s he ne wo k cos
depic ing h ee in e es ing zones: a i s a ea o o e -enginee ing, whe e he
mul iplexing e ec gene a ed is excessi e and he cos is no op imum; ano he one
o unde -enginee ing, whe e mul iplexing le els a e sca ce and he cos aises; and
inally one o good-enginee ing, whe e mul iplexing le els esul adequa e and
which co esponds o he op imum cos . Finally, cos s emain he same when
demand inc eases, mainly because o he high capaci y o he ibe -op ic links.
The cos s o deploymen is e y sensi i e o he deg ee o co e age ( he e minal nodes
by he ne wo k) and o he mul iplexing le el ( he ela ion be ween he numbe o
egions and he numbe o e minals). We he e o e checked he good-enginee ing zone
be o e de ining he hub s uc u e and de e mining he egion’s densi y. On he o he
22
hand, he deploymen opology is ex emely obus , handling g ea inc eases in demand
( igu e 4).
We used ou esul s ine- une and calib a e he algo i hms. We hen applied he ool o
he di e en geog aphical a eas in he Andalusian egion ( igu es 5 and 6, able 7).
Figu e 5: We show he esul s ob ained o he a ea o he Se ille p o ince. We
show he case o a design based on ou egions, a e checking ha his was he
case which co esponded o he good-enginee ing mul iplexing a ea. Fi s , in his
igu e we depic he physical ne wo k ha we p oposed o he p o ince o Se ille.
I e lec s a ing- ee s uc u e. The hin lines ep esen he ee ou line
co esponding o he in a egional communica ion sec ions and he hick line
ep esen s he ing ou line co esponding o in e egional communica ion.
23
Figu e 6: The logic design ne wo k o he p o ince o Se ille e lec s a ing-s a
s uc u e. The hin line o ms he s a ou line co esponding o he in a egional
communica ion sec ions and he hick line ep esen s he ing ou line
co esponding o in e egional communica ion. The g ey scales di e en ia e
among he ou egions o he Se ille p o ince. No e ha whi e zones a e non-
co e ed a eas, o example ag icul u al a eas.
The in es men equi ed o se up he ibe -op ic ne wo k was 62.21 million eu os. We
also ob ained a ios use ul o analyzing cos s wi h espec o ce ain pa ame e s, o
example se up cos wi h espec o he numbe o people eached (43.54 eu os pe
pe son), and wi h espec o he numbe o households (126.24 eu os pe household); he
la e is a be e index o economic e o , because i is households ha subsc ibe o
cable elecommunica ion con en s.
To al se up cos 62.21 million €
Pe cen o su i abili y g ade speci ica ions me 100 %
Cos o se up wi h espec o popula ion eached 43.54 €/inhabi an eached
Cos o se up wi h espec o he numbe o households
eached 126.74 €/household eached
24
Cos o se up wi h espec o he numbe o main esidences
eached
159.55 €/main household
eached
Table 7: Main economic a ios associa ed o he ne wo k se up in he a ea
co esponding o he Se ille p o ince. The mos in e es ing alues a e he o al
se up cos and he se up cos wi h espec o he numbe o households since hey
a e he subsc ibe s o he cable elecommunica ion con en s. Also, we ha e
included his a io wi h espec o he numbe o main esidences, aking in o
accoun he possibili y o ha ing se e al esidences pe household (as aca ion
esidence o example).
Conclusion
The di ec o a e gene al o he Andalusia Go e nmen o he in o ma ion socie y
ega ded as e y posi i e ou con ibu ion o he de elopmen o a ool o e alua ing
he le els o in es men i would equi e o se up a ibe -op ic ne wo k o he egion.
Ou esul s allowed i o jus i y i s in e es in es ablishing a ne wo k ha included small
owns as well as la ge ci ies.
The inal esul o ou wo k was a echnical s udy used as he basis o he cu en pilo
p ojec o he deploymen o a uni e sal elecommunica ion ne wo k in he Andalusia
egion. The egional go e nmen is cu en ly implemen ing he pilo p ojec and
ca ying ou he i s ci il-enginee ing wo ks. Fu he mo e, he p oposal is helping o
pe suade o p i a e ope a o s in Andalusia o he ad an ages o be de i ed om
es ablishing elecommunica ion capabili ies, especially in a ac ing he ea ly ma ke .
25