Vous êtes sur la page 1sur 32

2013

Analytic Hierarchy
Process
Solar energy company expands
abroad. Which country to
choose?
oat! Simina "alentina
#roup 13$
%ucharest &ni'ersity
o( )conomic Studies
*able o(
contents
+ntro
ducti
on
3
,ecision
problem -
Sur'ey $
,ata
processing
12
.onclusions
1/
oat!
Simin
a
"alen
tina 0
#roup
13$
+ntroductio
n
*he
sola
r
ene
rgy
co
mp
any
Adl
ers
ho(
bas
ed
in
#er
ma
ny
had
a
1ou
rish
ing
acti
'ity
in
the
last
dec
ade
2
part
icip
atin
g in
#er
ma
ny3
s
rep
utat
ion
o(
4orld4ide top
photo'oltaic
installer. As a
matter o( (act2
#ermany
established an
ob5ecti'e o(
producing 3$6
o( electricity
(rom rene4able
sources by
2020 and a
long7term goal
o( 1006 o(
electricity to be
generated (rom
green energy
by 20$0.
+n this sense2
the 'ision o(
Adlersho( is to
have a solar
plant on each
roof. 8or a
continuous and
sustainable
de'elopment2
the managing
team o( the
company
endea'ours to
expand the
business
abroad by
building solar
9elds and
in'esting in
energy
in(rastructure
4here'er they
decide to go.
*heir decision
is supported by
the (act that
sola
r
pan
els
inst
alla
tion
4ill
pea
:
the
se
yea
rs
in
#er
ma
ny
and
exp
and
ing
4ill
e'e
ntu
ally
bec
om
e
nec
ess
ary2
hen
ce
9rst
7
mo
'er
ad'
ant
age
abr
oad
is
o(
tre
me
ndo
us
imp
orta
nce due to
mar:et
penetration
and early
in'estments in
;,.
+n order to
approach the
issue o(
expansion2 the
company has to
establish
se'eral
alternati'e
countries and
a(ter4ards
choose the
most
appropriate o(
them. The
decision-
making
process
according to
the
alternatives
and criteria
available
represents
the objective
of the
present paper
and there will
be used the
Analytic
Hierarchy
Process. *he
steps are the
(ollo4ing<
State the
decision
problem
and create
the AHP
(rame4or:.
,e'elop
a sur'ey
to be
ans4ere
d by
minimu
m
2
int
ere
ste
d
par
ties
2
bas
ed
on
the
crit
eri
a
an
d
alt
ern
ati'
es
(ro
m
the
AH
P
str
uct
ure
.
Per
(or
m
pair
7
4is
e
co
mp
aris
ons
thr
ou
gh
the
scales<
=inguisti
c model2
Saaty
and >a
?heng.
Process
data in
decision
matrice
s2
establis
h the
priority
'ector
and
aggrega
te the
9nal
results
in #auss
10.
@bser'atio
ns and
conclusions
.
oat! Simina
"alentina 0
#roup 13$
,ecision problem
)xpanding a business abroad can be cumbersome because choosing
the country to conduct (uture acti'ities is in1uenced not only by the
business en'ironment and prospects in that particular country but also
by the potential outcomes i( another state 4ere to be chosen.
+n our case the main Auestion is<
Which country would be the most appropriate to set up an
Adlershof branch?
eaching an ans4er presupposes the creation o( an AHP structure that
con'eys the criteria and alternati'es in Auestion. *he possible countries (or
the upcoming expansion are< omania2 South A(rica and .hina. *he criteria
are represented by< energy in(rastructure2 secure legislati'e en'ironment and
cost eBecti'eness.
Problem .hoosing a
country
Criteria
Secure
legislati'e )nergy
.ost
eBecti'eness
en'ironment in(rastructure
South A(rica
Alternatives omania
.hina
AHP Decision Structure
oat! Simina "alentina 0 #roup 13$
nergy infrastructure
*his criterion re(ers to the capability o( the countries to support the
installation o( solar 9elds in terms o( designated areas (or solar energy. +n this
sense 4e may ha'e grids and transmission operators that are up to date 4ith
technology and are able to streamline the processes. +n addition2 energy
in(rastructure re(ers to a complex net4or: that (acilitates consumers3 access
to reliable electricity ser'ices. *he energy policies are desired to be compliant
4ith the installation procedures used by Adlersho(. #reen energy procedures
already underta:en Ce.g.< 4ind po4erD 4ould represent a plus (or Adlersho(2 a
starting point in cooperating 4ith other companies. ).g.< omania lac:s in
in(rastructure2 generally spea:ing2 there(ore initial in'estments are needed to
prepare the installation o( solar panels.
!ecure legislative environment
As the name already suggests2 besides the idea o( solar energy panels and
in(rastructure2 the la4 in that particular country should be supporti'e 4ith
international companies desiring to set up branches and by de(ault create
5obs and impro'e the standard o( li'ing. *he case may be that in countries
4ith a less democratic perspecti'e there is a general tendency to discourage
starting up business 4hen the shareholders are (oreigners. ).g.< .hina and
the contro'ersial past regarding the hostile ta:eo'er o( international
branches and closing se'eral others. *his is 'ery less probable no4adays2 but
a small Auestion mar: remains.
"ost #ectiveness
*his criterion re1ects the (uture prospects o( the company regarding the
labour costs2 material costs2 the costs o( the initial in'estments as 4ell as the
ability to become pro9table and increase pro9t margin in a relati'ely short
period o( time. ).g.< .hina is percei'ed as ha'ing the lo4est costs but there is
an increasing trend in 4or:ers 34ages and standard o( li'ing2 hereby
suggesting that South A(rica might be the Ene4F .hina2 although its
in(rastructure reAuires extensi'e impro'ements.
Sur'ey
*he sur'ey 4ill be ans4ered by a representati'e o( the managing team o(
Adlersho( denominated Respondent 1 and an expert in rene4able energy
(rom the )uropean )nergy .entre denominated Respondent 2. *he sur'ey is
aimed at supporting Adlersho( in ta:ing an appropriate decision.
oat! Simina "alentina 0 #roup 13$
Within the sur'ey2 there 4ill be used the Saaty3s scale2 4hich is a G7step
scale help(ul (or establishing the intensity of the preferences and (or avoiding
irrationality or sub5ecti'e in(erences. . +n addition2 the Saaty scale 4ill be
transposed in =inguistic and >a ?heng scales as 4ell. Ho4e'er2 the
computations corresponding to the matrices (or pair74ise comparisons used
in #auss 4ill be based on the Saaty scale.
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtreme
ly
importan
t more more plus more more plus more more strongly more
important important important
importan
t important important more
importa
nt
importan
t
Saaty 1 2 3 - $ H I / G
=inguisti
c
!
$
!
%
!
&
!
'
!
(
!
)
!
*
!
+
!
,
>a 1 GJ/ GJI GJH GJ$ GJ- GJ3 GJ2 G
?heng
!tages of the survey-
Pair 4ise comparison bet4een each 2 criteria (rom the 3 criteria in
terms o( country choice.
Pair 4ise comparison bet4een each 2 alternati'es (rom the 9rst
criterion perspecti'e Csecure legislati'e en'ironmentD
Pair 4ise comparison bet4een each 2 alternati'es (rom the second
criterion perspecti'e Cenergy in(rastructureD
Pair 4ise comparison bet4een each 2 alternati'es (rom the third
criterion perspecti'e Ccost eBecti'enessD
oat! Simina "alentina
0 #roup 13$
!urvey .uestions-
1. Which o( the criteria you consider to be more important?
1)Secure legislati'e en'ironment
2) )nergy in(rastructure
Pair 04ise comparison (or intensity e.g.< Ho4 much more important is the
chosen one compared to the others?
@n a scale (rom 1 to G C17eAually important2 G7extremely importantD by
ho4 much is more important in your decision the chosen criteria?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
importan
t more more plus more more plus more more strongly more
important important important important important important more
importan
t
important
Saaty 1 2 3 - $ H I / G
2. Which o( the criteria you consider to be more important?
1) Secure legislati'e en'ironment
2) .ost eBecti'eness
@n a scale (rom 1 to G C17eAually important2 G7extremely importantD by
ho4 much is more important in your decision the chosen criteria?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
importan
t more more plus more more plus more more strongly more
important important important important important important more
importan
t
important
Saaty 1 2 3 - $ H I / G
3. Which o( the criteria you consider to be more important?
1) )nergy +n(rastructure
2) .ost eBecti'eness
@n a scale (rom 1 to G C17eAually important2 G7extremely importantD by
ho4 much is more important in your decision the chosen criteria?
oat! Simina "alentina
0 #roup 13$
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
importan
t more more plus more more plus more more strongly more
important important important important important important more
importan
t
important
Saaty 1 2 3 - $ H I / G
4. Which country is more pre(erable (rom secure legislati'e en'ironment
'ie4point? aDomania
bD.hina
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
$. Which country is more pre(erable (rom secure legislati'e en'ironment 'ie4point?
aDomania
bDSouth
A(rica
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
H. Which country is more pre(erable (rom secure legislati'e
en'ironment 'ie4point? aD.hina
bDSouth A(rica
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
oat! Simina "alentina 0 #roup 13$
7. Which country is more pre(erable in terms o( cost eBecti'eness?
1)omania
2) .hina
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
8. Which country is more pre(erable in terms o( cost eBecti'eness?
1) South A(rica
2) .hina
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
9. Which country is more pre(erable in terms o( cost eBecti'eness?
1)South A(rica
2) omania
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
oat! Simina "alentina 0 #roup 13$
10. Which country is more pre(erable in terms o( energy in(rastructure?
1) omania
2) .hina
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
11. Which country is more pre(erable in terms o( energy in(rastructure?
1)omania
2) South A(rica
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
12. Which country is more pre(erable in terms o( energy in(rastructure?
1) South A(rica
2) .hina
@n a scale (rom 1 to G by ho4 much is more pre(erable your choice?
Scale )Aually Wea:ly
>oderatel
y >oderately Strongly Strongly
,emonstrate
d
"ery2
'ery
)xtremel
y
pre(erabl
e more more plus more more plus more more strongly more
pre(erable pre(erable pre(erable pre(erable
pre(erabl
e pre(erable more
pre(erabl
e
pre(erabl
e
Saaty 1 2 3 - $ H I / G
oat! Simina "alentina 0 #roup 13$
8urther on2 it is important to (ormaliKe the alternati'es and criteria
as (ollo4s<
Secure legislati'e en'ironment 0 .1
.ost eBecti'eness 0 .2
)nergy in(rastructure 0 .3
omania 0 A1
South A(rica 0 A2
.hina 0 A3
!urvey answers /espondent % 0using formali1ed form for
criteria and alternatives2-
1. .1 to .3 pre(erred 4ith G
2. .1 to .2 pre(erred 4ith /
3. .2 to .3 pre(erred 4ith H
4. A1 to A3 pre(erred 4ith I according to .1
5. A1 to A2 pre(erred 4ith 3 according to .1
6. A2 to A3 pre(erred 4ith - according to .1
7. A3 to A1 pre(erred 4ith / according to .2
8. A2 to A3 pre(erred 4ith 3 according to .2
9. A2 to A1 pre(erred 4ith G according to .2
10. A3 to A1 pre(erred 4ith $ according to .3
11. A1 to A2 pre(erred 4ith $ according to .3
12. A3 to A2 pre(erred 4ith / according to .3
!urvey answers /espondent & 0using formali1ed form for
criteria and alternatives2-
1. .1 to .3 pre(erred 4ith G
2. .1 to .2 pre(erred 4ith H
3. .2 to .3 pre(erred 4ith 3
4. A1 to A3 pre(erred 4ith G according to .1
5. A2 to A1 pre(erred 4ith 3 according to .1
6. A2 to A3 pre(erred 4ith - according to .1
7. A3 to A1 pre(erred 4ith H according to .2
8. A3 to A2 pre(erred 4ith - according to .2
9. A2 to A1 pre(erred 4ith G according to .2
10. A3 to A1 pre(erred 4ith - according to .3
11. A1 to A2 pre(erred 4ith 2 according to .3
12. A3 to A2 pre(erred 4ith / according to .3
oat! Simina
"alentina 0 #roup 13$
,ata processing
8urthermore2 the decision matrices 4ill be constructed (or Saaty2 =inguistic
and >a ?heng scales (or espondent 1 and espondent 2. *his is necessary
in order to determine the priority 'ectors aLliated to each decision matrix2
important in determining the synthesised priorities.
Respondent 1
!aaty
,>M. .1 .2 .3
.1 1 / G 41
.2 1J/ 1 H 42
.3 1JG 1JH 1 43
,>M.1 A1 A2 A3
A1 1 3 I a
A2 1J3 1 - b
A3 1JI 1J- 1 c
,>M.2 A1 A2 A3
A1 1 1JG 1J/ d
A2 G 1 3 e
A3 / 1J3 1 (
,>M.3 A1 A2 A3
A1 1 $ $ g
A2 1J$ 1 1J$ h
A3 1J/ / 1 i
Respon
dent 2
,>M. .1 .2
.1 1
.2 1JH
.3 1JG 1J3
,>M.1 A1 A2
A1 1 3
A2 1J3 1
A3 1JG N
,>M.2 A1 A2
A1 1 1JG
A2 G 1
A3 H -
,>M.3 A1 A2 A3
A1 1 2 1J- g
A2 1J2 1 1J/ h
A3 - / 1 i
oat! Simina
"alentina 0
#roup 13$
Respondent 1
3a
4heng
,>M. .1 .2 .3
.1 1 GJ2 G 41
.2 2JG 1 GJ- 42
.3 1JG -JG 1 43
Respondent 2
,>M. .1 .2 .3
.1 1 GJ- G 41
.2 -JG 1 GJI 42
.3 1JG IJG 1 43
,>M.1 A1 A2
A1 1 GJI
A2 IJG
A3 3JG HJG
,>M.1 A1 A2 A3
A1 1 GJI G a
A2 IJG 1 GJH b
A3 1JG HJG 1 c
,>M.2 A1
A1 1
A2 G
A3 GJ2
,>M.2 A1 A2 A3
A1 1 1JG -JG d
A2 G 1 HJG e
A3 GJ- GJH 1 (
,>M.3
A1
A2
A3
,>M.3 A1 A2 A3
A1 1 GJ/ HJG g
A2 /JG 1 2JG h
A3 GJH GJ2 1 i
oat!
Simina
"alenti
na 0
#roup
13$
Respondent 1
5inguist
ic
,>M. .1 .2 .3
.1 !
$
!
+
!
,
41
.2 !-
+
!
$
!
)
42
.3 !-
,
!-
)
!
$
43
,>M.1 A1 A2 A3
A1 !
$
!
&
!
*
a
A2 !-
&
!
$
!
'
b
A3 !-
*
!-
'
!
$
c
,>M.2 A1 A2 A3
A1 !
$
!-
,
!-
+
d
A2 !, !$ !& e
A3 !+ !-
&
!$ (
,>M.3 A1 A2 A3
A1 !
$
!-
,
!-
+
g
A2 !
,
!
$
!
&
h
A3 !
+
!-
&
!
$
i
Respondent 2
,>M. .1 .2 .3
.1 !
$
!
)
!
,
41
.2 !-
)
!
$
!
&
42
.3 !-
,
!-
&
!
$
43
,>M.1 A1 A2 A3
A1 !
$
!
&
!
,
a
A2 !-
&
!
$
!
'
b
A3 !-
,
!-
'
!
$
c
,>M.2 A1 A2 A3
A1 !
$
!-
,
!-
)
d
A2 !, !$ !-
'
e
A3 !
)
!
'
!
$
(
,>M.3 A1 A2 A3
A1 !
$
!-
,
!-
'
g
A2 !
,
!
$
!-
+
h
A3 !' !+ !$ i
oat! Simina
"alentina 0 #roup
13$
Priority 'ectors determined 4ith the use o( #auss program and the
corresponding line o( code<
new;
let DM_C[3,3]= the numbers from the matrix are introduced with one spacebar among
each other, following each line ;
print DM_C DM_C;
!"_DM_C=#eros$3,%&;
DM_C_scaled=#eros$3,3&;
n='((;
i=%;
do until i)n;
*=%;
do until *)rows$DM_C&;
DM_C_scaled[+,*]=DM_C[+,*]+,sumC$DM_C[+,*]&;
*=*-%;
endo;
DM_C_scaled;
DM_C_scaled.DM_C_scaled;
i=i-%;
endo;
!"_DM_C=DM_C_scaled[+,%];
print!"_DM_C !"_DM_C;
end;
"omputations of priority vectors according to !aaty !cale-
Respondent 1 Respondent 2
DM_C
oat! Simina
"alentina 0 #roup 13$
DM_C1
DM_C2
DM_C3
"omputations of priority vectors according to 3a
4heng !cale-
Respondent 1 Respondent 2
DM_C
oat! Simina
"alentina 0 #roup
13$
DM_C1
DM_C2
DM_C3
Areation of the priority !ectors accordin to the synthesi"ed
form #!ectors deri!ed from the $aaty $cale constructed matrices%&
C1 C2 C3 Respondent 1 Respondent 2
w1 w2 w3
A1 a d
aw1 ' dw2 '
w3
0.679*0.8
09 + 0.055*0.101
0.69*0.7
50 +0.062*0.12! +
+0.75!*0.089 " ()*21
0.181*0.1
2! "()+,-
A2 b e h
bw1 ' ew2 '
hw3
0.22!*0.809 +0.5*0.101
+0.150*0.089
0.229*0.7
50 +0.562*0.12! +
"()2,, 0.090*0.12! " ()2+1
A3 c f i
cw1 'fw2
'iw3
0.096*0.8
09 +0.!!*0.101 + 0.077*0.750 +0.75*0.12! +
0.09!*0.0
89 "()12.
0.727*0.1
2! "()1.3
oat! Simina "alentina
0 #roup 13$
.onclusio
ns
*he
syn
the
siK
ed
(or
m
o(
the
dec
isio
n
ma
tric
es
re'
eal
s
tha
t
the
pri
orit
y
'ec
tor
ass
oci
ate
d
to
A1
is
the
lar
ges
t2
me
ani
ng
tha
t
A1
Co
ma
nia
D
rep
res
ents the most
appropriate
alternati'e to
be considered
by the solar
energy
company.
Another
obser'ation
that can be
made is
related to the
beha'iour o(
the
respondents
4ith respect to
the
alternati'es
chosen. *he
analysis
concludes that
both o( them
4ould reach
the conclusion
that A1 is the
best choice (or
the company.
+( it had been a
contradiction
in the results
o( the analysis2
4e 4ould ha'e
per(ormed
computations
in order to
determine the
.onsistency
+ndex (or each
decision
matrix2 (or
both
respondents.
*he
respondent
4ho had the
most
con
sist
ent
ma
tric
es
or
a
lar
ger
nu
mb
er
o(
mo
re
con
sist
ent
ma
tric
es
4o
uld
ha'
e been ta:en
into account
(or an accurate
decision.
All things
considered2
the #erman
solar energy
company
should expand
its business in
omania and it
can be in(erred
that the secure
legislati'e
en'ironment
associated
4ith
omania3s
accession to
)& has played
a ma5or role in
the decision7
ma:ing
process.
oat
!
Simi
na
"ale
ntin
a 0
#ro
up
13$