Acta geographica Slovenica, 58-1, 2018, 97–108
MULTI-CRITERIA ASSESSMENT OF LESS
FAVOURED AREAS: A STATE LEVEL
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk,
Matej Kolenko, Črtomir Rozman
Less Favoured Areas (LFAs) where production conditions are difficult.
K
A
R
M
E
N
 P
A
Ž
E
K
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 97
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk, Matej Kolenko, Črtomir Rozman, Multi-criteria …
98
DOI: https://doi.org/10.3986/AGS.962
UDC: 332.6:631(497.4)
COBISS: 1.01
Multi-criteria assessment of less favoured areas: A state level
ABSTRACT: The paper present a multi-criteria decision DEXi model for assessment of less favoured areas
(LFAs). The tool enables easier assessment of farming in different areas of Slovene LFAs with respect to
criteria of sustainability. Analysis of LFAs and final integration of the assessment of LFAs depend upon
various criteria. In this paper we analyze individual LFAs and farming systems in these areas at the state
level with respect to criteria of sustainability and farming potential.
KEY WORDS: geography, less favoured areas, agricultural policy, multi-criteria decision analysis, DEXi
Večkriterijska ocena območij z omejenimi možnostmi za kmetijsko dejavnost: stanje
v državi
POVZETEK: V pris pev ku je pred stav ljen več kri te rij ski odlo či tve ni model DEX i za oce no obmo čij z omejeni -
mi možnostmi za kme tij sko dejav nost (OMD). Raz vi to orod je omo go ča oce no nači na kme to va nja s pou dar kom
na kri te ri ju trajnostnosti v raz lič nih OMD območ jih. Ana li za OMD obmo čij in model na inte gral na konč na
ocea na kaže ta na odvi snost več kri te ri jev. Pri oce ni posa mez ne ga OMD območ ja in nači na kme to va nja na
nivo ju Slo ve ni je ima ta tako pomemb no vlo go kri te rij traj no sti in poten cial posa mez ne ana li zi ra ne kme ti je.
KLJUČNE BESEDE: geo gra fi ja, območ ja z ome je ni mi možnostmi, kme tij ska poli ti ka, več kri te rij ska odloči -
tve na ana li za, DEX i
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk, Matej Kolenko, Črtomir Rozman
University of Maribor, Faculty of Agriculture and Life Sciences
karmen.pazek@um.si, ales.irgolic@um.si, jernej.turk@um.si, andreja.borec@um.si, jernej.prisenk@um.si,
matej.kolenko@gmail.com, crt.rozman@um.si
The paper was submitted for publication on October 9th, 2013.
Ured niš tvo je pre je lo pris pe vek 9. ok to bra 2013.
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 98
1 Introduction
The proportion of the total utilized agricultural area (UAA) classified as less favoured areas (LFAs) has
raised from 33% to 65% in European Union in the last two decades. A substantial amount of the utilized
agricultural area is classified as mountain areas (MacDonald etal. 2000). This situation is also seen in Slovenia,
where 491.000 hectares or 72.3% of the UAA are located in mountainous and hilly areas (LFAs). Almost
two-thirds of it is in permanent pasture, and arable land accounts for less than 30%. The proposed European
Union indicators for defining areas less suitable for agriculture (there are eight European criteria) in Slovenia
are not entirely appropriate because taking them into account would omit some distinctly and clearly unsuit-
able areas, i.e. karst areas (Ciglič et al. 2012). However, as other European Union Alpine regions, Slovenia
is characterized by one of the most difficult conditions for agricultural production in Europe. Overutilization
of agricultural land is becoming a serious problem, although the reasons for overutilization are not the
same everywhere. Efforts to combat overutilization and understanding of the background to the problem
are of particular importance for improvement of agricultural land quality (Borec et al. 2004). The current
rural development European policy includes significant evolution of support schemes for LFAs. Agricultural
production in LFAs is usually extensive and less suited for different farming systems and agri-food pro-
duction. Some authors have suggested development strategies for LFAs based on interdisciplinary research
of the coupling of human and natural systems approaches (Ruben et al. 2005). Sheate et al. (2008) examined
the sustainability of various scenarios for reconciling biodiversity conservation with declining agriculture
use in mountain areas of Europe. Their methodology was grounded in baseline of ecological and
socio-economic data. Terluin et al. (1995) examined agricultural incomes in LFAs from an economic per-
spective. In their scenario, income was based on the typology of European countries and the relationship
of regional gross domestic product per inhabitant and farm net value added per annual work unit. They
confirmed that within the analyzed geographical areas, farmers in LFAs receive a higher amount of direct
income subsidies than farmers in regions not classified as LFAs. The links between size, subsidies and per-
formance for Slovenian farms were presented by Bojnec and Latruffe (2013). The study concludes that
Slovenian farms have always been small and highly subsidized. Further, a conception of developmental
types of mountain is presented regarding the state of developmental potentials on farms by Kerbler (2003).
Assessment of farming potential on individual LFA is usually related to multiple criteria (Pažek et al. 2010).
Tiwari et al. (1999) asserted that the rural reality system is complex and that the use of economic or envi-
ronmental criteria alone may be insufficient. Multiple competitive criteria are likely to influence the
decision-making process. A decision model must be able to evaluate all the options when considering those
factors influencing the decision. A multi-criteria decision analysis approach was used in this paper to assess
different organizational and planning decisions in farm management, such as the DEXi methodology
(Bohanec et al. 1995; Bohanec, Zupan and Rajkovič 2000; Pažek et al. 2006; Rozman et al. 2006; Bohanec,
Džeroski and Žnidaršič, 2004; Bohanec et al. 2007; Pavlovič et al. 2011, Pažek, Rozman and Irgolič 2012a).
The theoretical background of hierarchical multi-attribute decision models is based on the dissection of
a complex decision problem into smaller and less complex subproblems. In this context, the DEXi method
uses qualitative variables and utility functions in the form of decision rules and provides qualitative assess-
ments of alternatives. Subproblems are represented by variables, which are organized into a hierarchy
(Pažek et al. 2012b).
The aim of this paper is to present a DEXi multi-criteria decision support tool for assessment of LFAs
and farming systems in these areas at the state level.
2 Materials and methods
The DEXi method is a combination of traditional multi-attribute decision-making processes and spe -
cific elements of expert system and machine learning techniques (Bohanec 2003). Computer program for
multi-attribute decision making is called DEXi. It is aimed at interactive development of qualitative
multi-attribute decision models and the evaluation of options. This is useful for supporting complex deci-
sion-making tasks, where there is a need to select a particular option from a set of possible ones so as to
satisfy the goals of the decision maker (Bohanec 2014). Variables are connected by utility functions. Utility
Acta geographica Slovenica, 58-1, 2018
99
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 99
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk, Matej Kolenko, Črtomir Rozman, Multi-criteria …
functions in DEXi are adjusted to qualitative variables and, therefore, represented by »if/then« decision
rules (elementary decision rules), which are usually given in a tabular form. The DEXi method can be used
for solving various decision problems regarding real-world decisions (Bohanec et al. 1995; Bohanec and
Rajkovič 1999; Bohanec, Zupan and Rajkovič 2000; Bohanec et al. 2006).
The DEXi models are developed by defining:
• attributes (a): qualitative variables that represent decision subproblems (for instance farm size, as demon-
strated in Table 1);
• scales: ordered or unordered sets of symbolic values that can be assigned to attributes (for instance for
farm size: small to 6.00 ha, average (between 6.00 and 7.00 ha) and big (farm size is over 7.00 ha; as demon-
strated in Table 2);
• tree of attributes: a hierarchical structure representing the decomposition of the decision problem;
• utility functions: rules that define the aggregation of attributes from bottom to the top of the tree of attrib-
utes.
In the evaluation and analysis stage, DEXi facilitates:
• description of options: defining the values of basic attributes (terminal nodes of the tree);
• evaluation of options: a bottom up aggregation of basic attributes values based on utility functions;
• analysis of options: »what-if« analysis, »plus-minus-1« analysis, selective explanation and comparison
of options, and
• reporting: graphical and textual presentation of models, options and evaluation results.
The hierarchical model structure for the assessment of LFAs that represents the decomposition of the
decision problem into subproblems, was defined by the policy decision maker expert group of the Agency
of the Republic of Slovenia for agricultural markets and rural development (AKTRP 2012; Irgolič 2011).
The main criteria included in the model structure are: description of the farm, farm holder age structure,
100
Table 1: Hierarchical model structure for assessment of LFAs.
Attribute (a)
Amount of protein crops payments and nuts payments
Amount of additional payments for milk
Amount of additional payments for beef
–
–
–
Final assessment
– Farm description
– Farm size
– Usage type of agricultural land
– Number of assigned points
– Amount of LFA payments
– Social structure
– Farm holder´s age
– Successor or not
– Successor´s age
– Amount of natural handicap payments
– Number of payment entitlements
– Value of payment entitlements for arable land
– Value of payment entitlements for pasture
– Amount of payment entitlements
– Amount of agri-environment payments
– Organic farming
– Integrated crop production
– Implementation of other agri-environmental measures
– Amount of payments
– Amount of direct payments
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 100
amount of natural handicap payments, amount of agri-environmental payments, and amount of direct pay-
ments. Aggregate criteria were divided into groups of criteria (as seen in Table 1) and in the final evaluation
of the LFAs areas: other areas, hill areas, karst areas, steep slopes, and mountain areas.
The attributes at the lowest level are basic descriptors of options (in our case individual LFAs), These
represent model inputs and must be provided by the decision maker. Table 2 presents the sets of scales
that were defined for all attributes in the model.
The decision rules are presented in a so-called complex form, with headings displaying approximate
weights assigned to the attributes (second row in Table 3). The so-called »weight-based strategy« of defin-
ing decision rules was used. In terms of the subattributes and the number of points assigned, if the value
of the assessment points by the agency for the observed farm was less than 310, this subattribute was assigned
the discrete value »bad« by the the DEXi model. If the farm was awarded between 311 and 350 points,
then the discrete value was »good.« Finally, if the farm received more than 351 points, the discrete value
assigned by the DEXi model was »excellent.« The symbols »≤«, »≥« define value intervals for the relevant
attribute. The asterisk »*« defines any possible value. The relative importance of the attributes was expressed
by weights (as seen at the top of Table 3). These weights were estimated by DEXi using a linear regression
method (according to Rozman et al. 2009), where DEXi interpolates the values of previously undefined
rules in the table. Linear coefficients respond to the required weights, and its surface lies as close as pos-
sible to the initially specific subset of rules (Pavlovič et al. 2011). In practical use, this means the higher
the weight, the more important the attribute.
After each attribute was assigned to a scale, the utility functions were defined (Table 3). The utility
functions evaluate and define individual attributes with respect to their immediate descendants in the hier-
archy. The utility function procedure was derived for each level in the hierarchy (partial utility function
for aggregate attributes and overall utility function for the whole model, except for the lowest level in the
Acta geographica Slovenica, 58-1, 2018
101
Amount of protein crops payments and nuts payments
Amount of additional payments for milk
Amount of additional payments for beef
–
–
–
(€)
(€)
(€)
Final assessment
– Farm description
– Farm size
– Usage type of agricultural land
– Number of assigned points
– Amount of LFA payments
(ha)
(points)
(€)
– Social structure
– Farm holder´s age
– Successor or not
– Successor´s age
(years)
(years)
– Amount of natural handicap payments
– Number of payment entitlements
– Value of payment entitlements for arable land
– Value of payment entitlements for pasture
– Amount of payment entitlements
(€)
(€)
(€)
– Amount of agri-environment payments
– Organic farming
– Integrated crop production
– Implementation of other agri-environmental measures
– Amount of payments
(%)
(%)
(%)
(€)
– Amount of direct payments
Table 2: Basic structure of the decision model, with sets of values (scales).
Attribute (a) Scale
INAPPROPRIATE; RATHER INAPPROPRIATE; APPROPRIATE; EXCELLENT
BAD; GOOD; EXCELLENT
SMALL; AVERAGE; BIG
Meadows; Plantations; Fields
< 310; 310–350; > 350
< 500; 500–1000; > 1000
BAD; GOOD
> 55; 40–55; 18–25; 25–40
No; Yes
> 55; 40–55; 18–25; 25–40
Bad; Good; Very good
< 6,5; 6,5–7,5; > 7,5
< 380; 380–400; > 400
< 160; 160–180; > 180
< 1000; 1000–1500; 1500–2000; > 2000
Bad; Good; Excellent
< 2; 2–4; 4–6; > 6
< 2; 2–4; 4–6; > 6
< 20; 20–30; 30–40; > 40
< 1200; 1200–1600; 160–2000; > 2000
Bad; Good
< 150; 150–250; > 250
< 500; 500–1000; > 1000
< 650; 650–750; > 750
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 101
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk, Matej Kolenko, Črtomir Rozman, Multi-criteria …
hierarchy). According to Pavlovič et al. (2011), for each attribute y, whose descendants in the hierarchi-
cal tree of attributes are x1, x2, … xn, the corresponding utility function f defines the mapping:
f: x1 × x2 × … xn → y,
where x1, x2, … xn and y denote values in the domains of the attributes a1, a2, … an and y. These rules define
the mapping of four subattributes and the assessment of the cumulative descriptive attributes of the farm
in the overall final assessment of the LFAs according to defined decision rules.
Table 3: Example of decision rules, with a utility function of the presented case.
Decision rules
Farm size Usage type Number of assigned Amount of LFA Farm
of agricultural land points payments description
26% 26% 31% 17%
1 SMALL Meadows < 310 * BAD
2 SMALL Meadows * < 500 BAD
3 SMALL * < 310 < 500 BAD
4 ≤ AVERAGE ≤ Plantations ≥ 310–350 ≥ 500–1000 GOOD
5 * ≤ Plantations 310–350 ≥ 500–1000 GOOD
6 ≤ AVERAGE Plantations * ≥ 500–1000 GOOD
7 * Plantations ≤ 310–350 ≥ 500–1000 GOOD
8 * ≥ Plantations < 310 ≥ 500–1000 GOOD
9 ≤ AVERAGE Plantations ≥ 310–350 * GOOD
10 ≤ AVERAGE ≥ Plantations ≥ 31–350 < 500 GOOD
11 * Plantations 310–350 * GOOD
12 * ≥ Plantations 310–350 < 500 GOOD
13 AVERAGE ≤ Plantations * * GOOD
14 AVERAGE * * < 500 GOOD
15 ≥ AVERAGE ≤ Plantations ≤ 310–350 * GOOD
16 ≥ AVERAGE * < 310 * GOOD
17 ≥ AVERAGE * ≤ 310–350 < 500 GOOD
18 * Fields ≥ 310–350 ≥ 500–1000 EXCELLENT
19 BIG * > 350 * EXCELLENT
The single line in Table 3, i.e., single decision rule, defines the value of the final assessment of LFAs
for one combination of values of the former four attributes. The description of the farm, amount of nat-
ural handicap, and agri-environmental payments can take three different discrete values (part of the decision
rules for the farm description attribute with the utility function is presented in Table 3), and the age struc-
ture and amount of direct payments can take two different values. Consequently, there are 108 possible
combinations (3 × 3 × 3 × 2 × 2) and thus 108 decision rules.
In the next step, the attribute values for each options were placed in the DEXi evaluation table, and
the evaluation analysis of the LFA assessment was evaluated.
Data for specific criteria compiled by the Agency of the Republic of Slovenia for agricultural markets
and rural development were used. The sample size was 42.856 Slovenian farms that are registered and finan-
cial supported by the government through different environmental programs, including support for farming
in LFA areas (D – 8.413 farms, H – 13.193 farms, K – 5.975 farms, S – 2.269 farms and V – 13.006 farms).
The average value of attributes from this database was used as input in the DEXi multi-criteria model (attrib-
utes at the leaves of the hierarchical tree as presented in Table 1).
102
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 102
Acta geographica Slovenica, 58-1, 2018
103
3 Result and discussion
The following Slovenian LFAs were included in the analysis (Irgolič 2011): hill areas (H), karst areas (K),
steep slopes (S), other areas (D), and mountain areas (V). Numerical and qualitative data compiled by the
Agency of the Republic of Slovenia for Agricultural Markets and Rural Development were employed. The
data were divided into five aggregate attributes (Table 4) according to the utility function, and an integrated
assessment of a particular LFA was performed at the end as the last step in the analysis.
The classification of particular areas was enabled by the multi-criteria decision model. The developed
DEXi model shows that the mountain areas were assigned a value of »excellent,« the best possible out-
come. As seen in the integrated assessment (Table 4 and Figure 1), most of the main attributes in this scenario
were assigned the highest value, except the attribute amount of agri-environmental payments, where the
utility function was assigned a value of »good.« This outcome is expected because there are more than 6%
of organic farms in mountain areas (awarded the highest discrete value) but less than 2% of integrated farms
(awarded the discrete value bad). Consequently, the implementation of other agri-environmental mea-
sures by those farms is between 30–40%, and the total amount of payments are €1.200–1.600/farm. Both
attributes intervals are assigned a neutral value in the model. The karst areas and steep slopes received the
worst assessments score (Figure 1 and Figure 2). Other areas and hill areas were evaluated as »appropri-
ate,« as seen in Table 4.
According to the lowest value assigned to two attributes in the assessment (amount of agri-environ-
mental and direct payments in the karst areas, where meadows predominate (33.133 ha meadows in
comparison with 9.791 ha of fields), the karst areas (K) results with the inappropriate final assessment. Steep
slopes (S) areas were assessed as »rather inappropriate.« In the hierarchical model, the farm description
attribute was assessed as »excellent.« Besides the subattribute »usage type of agricultural land« where mead-
ows predominate, all other subcriteria were assigned the highest discrete value.
In contrast to the karst areas, the neutral value (»very good«) was determined by the amount of nat-
ural handicap payments attribute. On the other site, two main attributes received the worst assessment
(»bad«), i.e., the amount of agri-environmental payments (weight 38%) and the amount of direct payments
(weight 11%).
The same integrated assessment was determined for the other areas and hill areas. In both alterna-
tives, the assessment of the attributes was the same. The description of farms in both areas was »excellent,«
the age structure criteria was assigned the value »good« (age of farm holder was between 40 and 55 years,
the farm has a successor), and amounts of both attributes were assessed as »good.«
The DEXi software also enables »what-if analysis.« For instance, one might consider how the overall
assessment can be improved. Table 5 shows the sensitivity (so called ± 1 analysis) for the areas K that was
originally assessed as »inappropriate«. The analysis shows, for example, which attributes considerably affect
the evaluation of areas K. When attribute Number of payment entitlements or Integrated crop produc-
tion increase the overall assessment of areas K improves to »rather inappropriate«.
The final assessment of the areas confirmed that selecting specific types of farming and other socio-eco-
nomic parameters in a particular LFA depends on various criteria, which were taken into consideration
in the multi-attribute decision model.
The developed model enabled a final assessment of the LFAs based on the defined attributes and the
decision rules within the defined utility functions for the observed problem. Moreover, the results show
that the developed decision model could be a suitable methodological tool to aid the practical evaluation
of different farming systems in LFAs and aid future political decision making. As this seems the use of pro-
gram desirable for many practical problems such as assessment of the service quality that embeds many
qualitative attributes or that cannot be easily numerically quantified further study should be particularly
focused on the integration of qualitative and quantitative modeling techniques in the assessment of ser-
vice quality as well as the inclusion of direct farm activities in the DEXi tree. However, despite the use of
qualitative data only, we found that the approach fulfilled most of our expectations and revealed considerable
advantages in comparison with other approaches. In particular, we emphasize the use of the qualitative
multi-criteria DEXi model, which was suitable in a field where judgment prevails, thus making it diffi-
cult to give numeric answers. This kind of model is comprehensible to a wide range of users in the evaluation
process.
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 103
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk, Matej Kolenko, Črtomir Rozman, Multi-criteria …
104
Am
ou
nt
 o
f p
ro
te
in
 cr
op
s p
ay
m
en
ts
 a
nd
 n
ut
s p
ay
m
en
ts
Am
ou
nt
 o
f a
dd
iti
on
al
 p
ay
m
en
ts
 fo
r m
ilk
Am
ou
nt
 o
f a
dd
iti
on
al
 p
ay
m
en
ts
 fo
r b
ee
f
– – –
(€
)
(€
)
(€
)
Fi
na
l a
ss
es
sm
en
t
–
Fa
rm
 d
es
cr
ip
tio
n
–
Fa
rm
 si
ze
–
Us
ag
e 
ty
pe
 o
f a
gr
icu
ltu
ra
l l
an
d
–
Nu
m
be
r o
f a
ss
ig
ne
d 
po
in
ts
–
Am
ou
nt
 o
f L
FA
 p
ay
m
en
ts
(h
a)
(p
oi
nt
s)
(€
)
–
So
ci
al
 st
ru
ct
ur
e
–
Fa
rm
 h
ol
de
r´
s a
ge
–
Su
cc
es
so
r o
r n
ot
–
Su
cc
es
so
r´
s a
ge
(y
ea
rs
)
(y
ea
rs
)
–
Am
ou
nt
 o
f n
at
ur
al
 h
an
di
ca
p 
pa
ym
en
ts
–
Nu
m
be
r o
f p
ay
m
en
t e
nt
itl
em
en
ts
–
Va
lu
e 
of
 p
ay
m
en
t e
nt
itl
em
en
ts
 fo
r a
ra
bl
e 
la
nd
–
Va
lu
e 
of
 p
ay
m
en
t e
nt
itl
em
en
ts
 fo
r p
as
tu
re
–
Am
ou
nt
 o
f p
ay
m
en
t e
nt
itl
em
en
ts
(€
)
(€
)
(€
)
–
Am
ou
nt
 o
f a
gr
i-
en
vi
ro
nm
en
t p
ay
m
en
ts
–
Or
ga
ni
c f
ar
m
in
g
–
In
te
gr
at
ed
 cr
op
 p
ro
du
ct
io
n
–
Im
pl
em
en
ta
tio
n 
of
 o
th
er
 a
gr
i-
en
vi
ro
nm
en
ta
l m
ea
su
re
s
–
Am
ou
nt
 o
f p
ay
m
en
ts
(%
)
(%
)
(%
)
(€
)
–
Am
ou
nt
 o
f d
ire
ct
 p
ay
m
en
ts
Ta
ble
 4:
 In
teg
rat
ed
 fin
al 
as
se
ssm
en
t o
f L
FA
s.
At
tr
ib
ut
e 
(a
)
EV
AL
UA
TI
ON
 R
ES
UL
TS
D
H
K
S
V
AP
PR
OP
RIA
TE
AP
PR
OP
RIA
TE
IN
AP
PR
OP
RIA
TE
RA
TH
ER
 IN
AP
PR
OP
RIA
TE
EX
CE
LL
EN
T
EX
CE
LL
EN
T
EX
CE
LL
EN
T
GO
OD
EX
CE
LL
EN
T
EX
CE
LL
EN
T
AV
ER
AG
E
SM
AL
L
AV
ER
AG
E
BIG
BIG
Fie
lds
Fie
lds
M
ea
do
ws
M
ea
do
ws
M
ea
do
ws
>
35
0
31
0–
35
0
<
31
0
>
35
0
>
35
0
50
0–
10
00
50
0–
10
00
>
10
00
>
10
00
>
10
00
GO
OD
GO
OD
GO
OD
GO
OD
GO
OD
40
–5
5
40
–5
5
40
–5
5
40
–5
5
40
–5
5
Ye
s
Ye
s
Ye
s
Ye
s
Ye
s
40
–5
5
40
–5
5
40
–5
5
40
–5
5
40
–5
5
Go
od
Go
od
Go
od
Ve
ry
 go
od
Ve
ry
 go
od
6,5
–7
,5
<
6,5
6,5
–7
,5
>
7,5
>
7,5
<
38
0
>
40
0
38
0–
40
0
>
40
0
>
40
0
<
16
0
>
18
0
16
0–
18
0
>
18
0
>
18
0
>
20
00
10
00
–1
50
0
15
00
–2
00
0
>
20
00
15
00
–2
00
0
Go
od
Go
od
Ba
d
Ba
d
Go
od
<
2
2–
4
2–
4
2–
4
>
6
>
6
2–
4
<
2
<
2
<
2
20
–3
0
20
–3
0
<
20
20
–3
0
30
–4
0
>
20
00
<
12
00
12
00
–1
60
0
12
00
–1
60
0
12
00
–1
60
0
Ba
d
Ba
d
Ba
d
Ba
d
Go
od
15
0–
25
0
>
25
0
15
0–
25
0
<
15
0
15
0–
25
0
<
50
0
<
50
0
<
50
0
50
0–
10
00
>
10
00
>
75
0
<
65
0
<
65
0
>
75
0
65
0–
75
0
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 104
Acta geographica Slovenica, 58-1, 2018
105
Good
Amount of direct payments
Amount of agri-environmental
payments
Final assessment
Amount of natural handicap payments
Social structure
Farm description
Good
Good
Amount of direct payments
Amount of agri-environmental
payments
Final assessment
Amount of natural handicap payments
Social structure
Farm description
Figure 1: Graphical presentation of the final assessment mountain areas (V)/gafični prikaz končne ocene gorsko-višinskih območij.
Figure 2: Graphical presentation of the final assessment of the karst (K) areas.
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 105
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk, Matej Kolenko, Črtomir Rozman, Multi-criteria …
4 Conclusion
According to the results the multi – attribute DEXi model can be regarded and applied practically to small-
er number of farms as well as to a broader sphere of scientifically research work. The developed model
may also be used in further development of agricultural policy, it also can be upgraded with the latest infor-
mation and adopted it to specific requirements. The multi – criteria methodology cannot replace or exclude
the policy decision maker experts but can serve as an additional instrument that enables faster analysis.
The model can be good basis and support tool for further development of more complex models that are
designed primarily for planning and decision-making process in agricultural policy especially by defini-
tion of different payments types in agriculture.
ACKNOWLEDGEMENTS: We express our grateful thanks to the anonymous reviewers and editor in chief
for the useful suggestion to improve the paper.
6 References
AKTRP 2012: Agency of the Republic of Slovenia for agricultural markets and rural development. Internet:
http://www.arsktrp.gov.si/en (20. 12. 2012).
Bohanec, M.  2014: DEXi. A  Program for multi-attribute decision making, version 4.00. Internet:
http://www-ai.ijs.si/MarkoBohanec/dexi.html (5. 11. 2014).
106
Amount of protein crops payments and nuts payments
Amount of additional payments for milk
Amount of additional payments for beef
–
–
–
(€)
(€)
(€)
Final assessment
– Farm description
– Farm size
– Usage type of agricultural land
– Number of assigned points
– Amount of LFA payments
(ha)
(points)
(€)
– Social structure
– Farm holder´s age
– Successor or not
– Successor´s age
(years)
(years)
– Amount of natural handicap payments
– Number of payment entitlements
– Value of payment entitlements for arable land
– Value of payment entitlements for pasture
– Amount of payment entitlements
(€)
(€)
(€)
– Amount of agri-environment payments
– Organic farming
– Integrated crop production
– Implementation of other agri-environmental measures
– Amount of payments
(%)
(%)
(%)
(€)
– Amount of direct payments
Table 5: The sensitivity analysis (± 1) for the karst areas (K).
Attribute (a) Plus-minus-1 analysis
–1 K +1
INAPPROPRIATE
AVERAGE
[ Meadows
[ < 310
> 1000 ]
40–55
yes ]
40–55
6,5–7,5 RATHER INAPPROPRIATE
380–400
160–180
1500–2000
2–4
[ < 2 RATHER INAPPROPRIATE
[ < 20
1200–1600
150–250
[ < 500
[ < 650
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 106
Bohanec, M. 2003: Decision support. Data mining and decision support: integration and collaboration.
Berlin.
Bohanec, M., Cortet, J., Griffiths, B., Žnidarsic, M., Debeljak, M., Caul, S., Thompson, J., Krogh, P. H. 2007:
A  qualitative multi-attribute model for assessing the impact of cropping systems on soil quality.
Pedobiologia 51-3. DOI: http://dx.doi.org/10.1016/j.pedobi.2007.03.006.
Bohanec, M., Džeroski, S., Žnidaršič, M. 2004: Multi-attribute modeling of economic and ecological impacts
of cropping systems. Informatica 28.
Bohanec, M., Messean, A., Scatasta, S., Angevin, F., Griffiths, B., Krogh, P.H., Žnidaršič, M., Džeroski, S. 2008:
A qualitative multi-attribute model for economic and ecological assessment of genetically modified
crops. Ecological Modelling 215, 1–3. DOI: http://dx.doi.org/10.1016/j.ecolmodel.2008.02.016
Bohanec, M., Messéan, A., Angevin, F., Žnidaršič, M. 2006: SMAC Advisor: A decision-support tool on
coexistence of genetically-modified and conventional maize. Proceedings information society IS 2006.
Ljubljana.
Bohanec, M, Rajkovič, V. 1999: Multi – attribute decision modeling: Industrial applications of DEX.
Informatica 23.
Bohanec, M., Rajkovič, V., Semolič, B., Pogačnik, A. 1995: Knowledge-based portfolio analysis for pro-
ject evaluation. Information and management 28. DOI: http://dx.doi.org/10.1016/0378-7206(94)00048-N.
Bohanec, M., Zupan, B., Rajkovič, V. 2000: Applications of qualitative multi-attribute decision models
in health care. International journal of medical informatics 58-59. DOI: http://dx.doi.org/10.1016/
S1386-5056(00)00087-3.
Borec, A., Flambart, A., Pažek, K. 2004: Relationships between production system of Slovenian mountain
farms and dynamics of overgrowing areas. Agricultura 3-1.
Bojnec, Š., Latruffe, L. 2013: Farm size, agricultural subsidies and farm performance in Slovenia. Land use
policy 32: 207-217. DOI: http://dx.doi.org/10.1016/j.landusepol.2012.09.016.
Ciglič, R., Hrvatin, M., Komac, B., Perko, D. 2012: Karst as a criterion for defining areas less suitable for
agriculture. Acta geographica Slovenica 52-1. DOI: http://dx.doi.org/10.1016/j.envsci.2010.08.003
Irgolič, A. 2011: Razvoj večkriterijskega modela za ocenjevanje območij z omejenimi dejavniki za kmetijsko
proizvodnjo. Magistrska naloga, Fakulteta za kmetijstvo in biosistemske vede Univerze v Mariboru.
Maribor.
Kerbler, B. 2003: A Conception of Developmental Typology of Mountain Farms: A Case Study of the
Municipality Ribnica na Pohorju. Acta geographica Slovenica 43-2. DOI: http://dx.doi.org/10.3986/
AGS43203.
MacDonald, D., Crabtree, J.R., Wiesinger, G., Dax, T., Stamou, N., Fleury, P., Gutierrez-Lazpita, J., Gibon, A. 2000:
Agricultural abandonment in mountain areas of Europe: Environmental consequences and policy response.
Journal of environmental management 59-1. DOI: http://dx.doi.org/10.1006/jema.1999.0335.
Pavlovič, M., Čerenak, A., Pavlovič, V., Rozman, Č., Pažek, K., Bohanec, M. 2011: Development of DEX-HOP
multi-attribute decision model for preliminary hop hybrids assessment. Computers and electronics
in agriculture 75-1. DOI: http://dx.doi.org/10.1016/j.compag.2010.11.002.
Pažek, K., Rozman, Č., Borec, A., Turk, J., Majkovič, D., Bavec, M., Bavec, F. 2006: The use of multi crite-
ria models for decision support on organic farms. Biological agriculture and horticulture 24-1. DOI:
http://dx.doi.org/10.1080/01448765.2006.9755009.
Pažek, K., Rozman, Č., Irgolič, A. 2012a: Ocena območij z omejenimi dejavniki za kmetijstvo z metodo
DEX. Maribor.
Pažek, K., Rozman, Č., Irgolič, A., Turk, J. 2012b: Multicriteria decision model for evaluating less favoured
areas for agricultural production. 47th Croatian & 7th international symposium on agriculture, 13–17 February,
Opatija, Croatia.
Rozman, Č., Potočnik, M., Pažek, K., Borec, A., Majkovič, D., Bohanec. M. 2009: A Multi-criteria assessment
of tourist farm service quality. Tourism management 30. DOI: http://dx.doi.org/10.1016/j.tourman.2008.11.008.
Pažek, K., Rozman, Č., Bavec, F., Borec, A., Bavec, M. 2010: A multi-criteria decision analysis framework
tool for the selection of farm business models on organic mountain farms. Journal of sustainable agri-
culture 34-7. DOI: http://dx.doi.org/10.1080/10440046.2010.507531.
Rozman, Č., Pažek, K., Bavec, M., Bavec, F., Turk, J., Majkovič, D. 2006: The Multi-criteria analysis of spelt
food processing alternatives on small organic farms. Journal of sustainable agriculture 28-2. DOI:
http://dx.doi.org/10.1300/J064v28n02_12.
Acta geographica Slovenica, 58-1, 2018
107
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 107
Karmen Pažek, Aleš Irgolič, Jernej Turk, Andreja Borec, Jernej Prišenk, Matej Kolenko, Črtomir Rozman, Multi-criteria …
Ruben, R., Kuiper, M.H., Pender, J. 2005: Searching development strategies for less-favoured areas. Netherlands
journal of agricultural sciences 53, 3–4. DOI: http://dx.doi.org/10.1016/S1573-5214(06)80012-6.
Sheate, W. R., Partidario, M. R., Byron, H., Bina, O., Dagg, S. 2008: Sustainability assessment of future sce-
narios: methodology and application to mountain areas of Europe. Environmental management 41.
DOI: http://dx.doi.org/10.1007/s00267-007-9051-9.
Terluin, I. J., Godeschalk, F. E., Meyer, H., Post, J. H., Strijker, D. 1995: Agricultural income in less favoured
areas of the EC: a regional approach. Journal of rural studies 11-2. DOI: http://dx.doi.org/10.1016/
0743-0167(95)00012-C.
Tiwari, D. N., Loof, R., Paudyal, G. N. 1999: Environmental-economic decision-making in lowland irrigated
agriculture using multi-objectives analysis techniques. Agricultural systems 60. DOI: http://dx.doi.org/
10.1016/S0308-521X(99)00021-9.
108
58-1-Special issue_08p_962-Karmen Pazek_acta49-1.qxd  12.9.2017  7:58  Page 108