11979


FACTA UNIVERSITATIS  

Series: Mechanical Engineering  

https://doi.org/10.22190/FUME230614026B 

© 2020 by University of Niš, Serbia | Creative Commons License: CC BY-NC-ND 

Original scientific paper 

APPLICATION OF THE DIBR II – ROUGH MABAC  

DECISION-MAKING MODEL FOR RANKING METHODS  

AND TECHNIQUES OF LEAN ORGANIZATION SYSTEMS 

MANAGEMENT IN THE PROCESS OF 

 TECHNICAL MAINTENANCE 

Darko Božanić1, Igor Epler1, Adis Puška2, Sanjib Biswas3,  

Dragan Marinković4,5, Stefan Koprivica5 

1Military Academy, University of Defence in Belgrade, Belgrade, Serbia 
2Department of Public Safety, Government of Brčko District of Bosnia and Herzegovina, 

Brčko, Bosnia and Herzegovina 
3Decision Science & Operations Management Area, Calcutta Business School, Diamond 

Harbour Road, West Bengal, India 
4Faculty of Mechanical Engineering, University of Niš, Niš, Serbia 

5TU Berlin, Department of Structural Analysis, Berlin, Germany 
6Faculty of Construction Management, Union - Nikola Tesla University, Belgrade, Serbia 

Abstract.  This paper presents a multi-criteria decision-making model based on the 

application of two methods, DIBR II and MABAC. The DIBR II method was used to 

define weight coefficients. The MABAC method was used to rank alternatives, and it 

was applied in a rough environment. Four experts were engaged in defining the 

criteria and alternatives as well as in the relation of criteria. The model was applied 

for ranking the methods and techniques of Lean organization systems management in 

the maintenance of technical systems of special purposes. At the end of the application 

was conducted a sensitivity analysis which proved the stability of the obtained results. 

Key words: Defining Interrelationships Between Ranked criteria II (DIBR II), Multi-

Attributive Border Approximation area Comparison (MABAC), Rough 

number, Lean concept  

                                                           
Received: June 14, 2023 / Accepted August 11, 2023 

Corresponding author: Darko Božanić  

Military Academy, University of Defence in Belgrade, Veljka Lukica Kurjaka 33, 11040, Belgrade, Serbia 

E-mail: darko.bozanic@va.mod.gov.rs     



2 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

  1. INTRODUCTION 

By analyzing basic technological process of technical maintenance (TM) of special 

purpose technical systems (SPTS) in a real maintenance workshop in the Army of Serbia, 

and its documentation which often lacks the data related to the organization of the process 

itself, time gaps were observed between the planned (normed) times of TM cycle of SPTS 

and the realized times of the SPTS technical maintenance cycle [1]. The periods of time 

in which SPTS wait for operation (considering administration and logistics) are generally 

not recorded in the workshop documentation, which makes it difficult to discuss about 

with the purpose of their reduction or total elimination.  

Many SPTS TM processes are not controlled, causing constant losses in the execution 

of TM on SPTS. Existing technological TM documentation for the most of the SPTS is 

“poor” regarding maintenance process management. For that reason, preparing computer 

data for software information system is very difficult as it relies only on currently 

available information from the TM technological process. 

Obviously, under such circumstances the management of SPTS TM processes must be 

fundamentally changed, primarily in terms of intensifying research and development of 

practical methods and techniques based on new concepts of maintenance, which are 

applied and constantly improved by leading companies in the neighborhood and in the 

world [2]. However, this is easier said than done. There are many difficulties in the 

management of TM of SPTS, and the most frequent and influential are as follows: poor 
quality and reliability of input data primarily contained in technological documentation; 

complexity of the maintenance system (CoM) of SPTS; many years of lack of investment 

in the management of TM, lack of exchange of work technologies, knowledge and 

innovations and inadequate training of top management for TM of SPTS to make them 

able to recognize and decide what is necessary and important to change in the 

maintenance management of SPTS so as to minimize losses. 

By eliminating negative effects of the mentioned factors, based on Lean organization 

systems management, current state of the work process and organization of TM of SPTS 

would probably change qualitatively. Qualitative changes would be reflected in [3]: 

-  shortening of SPTS TM cycle, 

-  optimal use of available capacities (spatial, human, material), 

-  reduction of costs per every maintained SPTS (increasing efficiency) and  

-  increased effectiveness of the TM process of SPTS. 
However, the application of all methods and techniques of Lean organization systems 

management requires significant investments. Considering this, a model for evaluation 

and ranking of the methods and techniques of Lean concept was developed. Depending on 

available resources, these methods and techniques (alternatives) would be applied in 

practice, primarily the first-ranked ones, and when available resources allow, the other 

alternatives.   

Previous research in the field of organization systems management [4,5] have proven 

positive effects of applying the Lean concept [6,7] in terms of the reduction of resources, 

both human and material, respectively increasing effectiveness and efficiency. There are 

indications that the application of the methods and tools of Lean organization systems 

management [8,9] can contribute to the improvement of current situation considering 

CoM of SPTS. The problem of applying Lean organization systems management has been 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 3 

discussed in several studies. In the following text, a part of the published research is 

analyzed. 

Tamanna and Rahman [10], based on the data of periods of operations until failure, 

criticality and consequences of failure of key equipment for maintenance of SIMGA1 

shipyard, wanted to analyze and select the optimal ship maintenance strategy with the 

goal of increasing the reliability and availability of the maintenance equipment and tools, 

reducing operating and production costs and increasing the shipyard's competitiveness.  

They observed the opportunity for improvement of shipyard maintenance in the 

application of Lean management tools. By using the AHP method, it was made a selection 

of the most effective Lean tools out of ten available.  

Singha Mahapatra and Shenoy [11] indicate the constant need of organizations for 

maintenance and services providing the elimination of activities and processes that do not 

contribute to the creation of new value for the customer or user, respectively, to the 

reduction of the price of the product or service. The authors state that the most managers 

of maintenance systems practically apply the tools of Lean management in different ways, 

because these proved satisfactory improvements in a relatively short period of use. 

Through their study, the authors strive to improve the methodology of selecting and 

applying Lean tools by identifying unique factors consisting the basis for evaluation of the 

existing rationality of maintenance in any maintenance organization, and which are the 

basis for measuring improvements after the application of Lean tools in a specific 

organization. 

Bhebhe and Zincume [12] analyzing systematically broad literature and applying 

general scientific method of deduction concluded that correctly dimensioned, constructed, 

properly exploited transport network with qualified managers was the key of economic 

power of every state. Based on the same data sources, they claimed that the most 

countries improved their economies by investing in transportation sector by identifying 

unnecessary costs generators, respectively, the losses within the maintenance function in 

transport companies. Observed cost generators were eliminated or reduced by systematic 

selection and application of Lean management tools that had already given the best results 

in the similar systems around the world. The paper provides guidelines for the 

improvement of overall condition by reducing unnecessary costs within transportation 

systems maintenance department, but also points out the opportunities for adjustment of 

application methods and its dynamics taking into account the specificities of as many 

transport companies as possible.  

Korchagin et al. [13] state that nowadays aircraft manufacturers around the world 

encounter the need and possibility for improvement of position and competitiveness of the 

aircraft manufacturing industry after the sale of the aircrafts. They quote that for hitting 

this goal the crucial thing is the improvement of the aircraft maintenance organization 

within all the phases of its lifespan by using the concepts such as Lean and Industry 4.0. 

In this paper was performed the integration of Lean management and Industry 4.0 

approaches, for the purpose of aircrafts maintenance, their modeling and simulation of 

their joint impact on the improvement of aircraft maintenance in the AniLogic simulation 

system. The results of the simulation showed that the simultaneous use of Lean 

management tools and Industry 4.0 from the beginning of the aircraft exploiting life in 

aviation industry would provide good results which would reflect in the aircraft 

maintenance efficiency increase, through the identification of bottlenecks in the process 



4 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

and making the right decisions in the direction of constant management of the 

maintenance process quality. 

Dragone et al. [14] in their paper emphasize that during the lifetime of the residential 

and commercial buildings are not carried out all the necessary maintenance activities 

often, which can cause serious damage and accidents. They consider that the maintenance 

of facilities is essential for the effectiveness of their utility value, as well as for the safety 

of the facilities. Traditional concepts of building maintenance management became 

ineffective over time, and do not provide good results because these are not oriented 

towards the processes and increasingly complex users’ demands and of the objects of 

work themselves – buildings. The authors deem that Lean maintenance management, with 

its principles and methods, can provide good solutions, applicable in the practice of 

facility maintenance.  

One of the most important parts of any decision-making process is the selection of the 

methods to be used [15-18]. Of course, additional dilemmas in these processes arise when 

choosing a possible way to modify standard methods, and for the purpose of better 

treatment of uncertainty in the decision-making process [19-22]. Through the analysis of 

the available studies, and taking into account the nature of the problem, two methods of 

multi-criteria decision-making were used in this research. The first method is Defining 

Interrelationships between Ranked criteria II (DIBR II), which is used for obtaining 

weight coefficients of criteria. This method is shown through the process of group 

decision-making. The DIBR II method is presented only in the study by [23] and it has 

not been used so far for solving complex decision-making problems. However, the 

simplicity that this method provides in communication with experts, as well as the simple 

mathematical model, recommended this method for solving the mentioned problem. The 

second method used in ranking of alternatives is MABAC, which proved as a very 

reliable one in up to date conducted researches. Considering the values, which were 

assigned to the alternatives at the beginning of the decision-making process when these 

were evaluated according to the criteria, the MABAC method was applied in a rough 

environment. 

Rough numbers were combined with decision-making methods in a large number of 

studies. Badi and Abdulshahed [24] used rough numbers to modify the AHP method, 

while in the study by [25] was presented the combination of rough numbers with 

DEMATEL method. Qi et al. [26] combine rough numbers with VIKOR method, Song et 

al. [27] with TOPSIS method, and Arsić et al. [28] with MAIRCA method. Rough 

numbers are commonly used in combination with fuzzy numbers, as in the studies by [29-

31]. The MABAC method is modified in literature by using rough numbers in multiple 

studies as [32-36]. 

Through the analysis up so far, two main contributions of this paper stand out. The 

first one is the application of the DIBR method in multi-criteria group decision-making, 

and for the first time as a part of the process of solving a real problem. The second 

contribution of the paper is related to the solving of a case study, respectively, the 

problem of ranking the methods and techniques of Lean concept in order to improve the 

management of the work process and the organization of the TM of SPTS in the Army of 

Serbia. 

The paper consists of several parts. In the second part the applied methods are 

described. Through the third part, that is, the case study, the definition and calculation of 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 5 

the weight coefficients of the criteria and the ranking of alternative solutions are carried 

out. The fourth part deals with sensitivity analysis, and at the end of the paper is provided 

the conclusion of this research. 

2. DESCRIPTION OF THE APPLIED METHODS  

This section describes the methods of DIBR II rough MABAC model. The phases and 

steps of the model are presented in the Fig. 1. 

Phase 1: Defining the criteria and calculation of weight 

coefficients of criteria using DIBR II method

Step 1. Definition of criteria

Step 2. Ranking criteria by their importance/weight

Step 3. Defining significance ratio between criteria

Step 4. Calculation of the relation of the most significant 

            criterion with other criteria

Step 5. Calculation of the value of the most significant 

            criterion

Step 6. Calculation of weight coefficients of other criteria

Step 7. Evaluation of quality of defined significance of the

            adjacent criteria

Step 8. Aggregation of weight coefficients of criteria for 

            each expert

Phase 2: Identification and ranking of alternatives 

using rough MABAC method

Step 1. Forming of initial decision-making matrix 

Step 2. Normalization of the elements of the initial

            decision-making matrix

Step 3. The calculation of the elements of the weighted 

            matrix 

Step 4. Defining border approximation area matrix

Step 5. Calculation of the matrix elements for alternatives 

            distance from the BAA 

Step 6. Ranking of alternatives

Step 7. Converting a rough number into the crisp value

Phase 3: Sensitivity analysis through the changes of weight coefficients of criteria 

 

Fig. 1 DIBR II – rough MABAC model  

As it can be observed from the Fig. 1, the model has three phases, where the criteria 

and their weight coefficients are defined first. Next, in the second phase of the model, the 

selection of the best alternative is made and at the end, it is checked the sensitivity of the 

model.  

2.1 Defining Interrelationships Between Ranked Criteria II Method 

The DIBR II method is developed by Božanić and Pamučar [23]. The method was 

developed for defining weight coefficients of criteria and consists of eight steps presented 

below. 

Step 1. Definition of criteria. As a part of the first step, a set of criteria is defined 

C={C1, C2,...Cn}, based on which the alternative solutions are ranked. 

Step 2. Ranking criteria by their importance/weight. All the criteria in the set C are 

ranked from the most to the least significant. For simple presentation of the method, the 

set of criteria is defined so that the criterion 1C  is the most significant, while the criterion 

Cn is the least significant, respectively, it is defined as follows C1>C2>C3>...>Cn. 

Step 3. Defining significance ratio between criteria.  



6 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

Step 3.1. Defining significance ratio between adjacent criteria. For each two adjacent 

criteria, their significance ratio is defined (ηj,j+1 where  , 1 1,2 2,3 3,4 1,, , ,...,j j n n      ). 
Thus, for example, for the comparison of the criteria C1 and C2 the significance ratio η1,2 

is defined. For this significance ratio it is valid that ηj,j+1≥1. The value ηj,j+1 shows how 

much more significant the criterion Cj is than the criterion Cj+1. According to everything 

stated, the following relations are defined through this step: 

 11 2 1,2 1,2
2

: :1
w

w w
w

     (1) 

 22 3 2,3 2,3
3

: :1
w

w w
w

     (2) 

 ...   

 
1

1 1, 1,: :1
n

n n n n n n
n

w
w w

w
       (3) 

Step 3.2. Defining significance relations between the most and the least significant 

criteria. In this step, the following relation is defined: 

 11 1, 1,: :1n n n
n

w
w w

w
     (4) 

This relation has the role of the control factor in evaluating the quality of the defined 

significance of the adjacent criteria.  

Step 4. Calculation of the relation of the most significant criterion with other criteria. 

Based on Eqs. (1-3), the value of the second and the other lower in range criteria is 

presented through the most significant criterion, as follows: 

- Based on Eq. (1), the value of the weight coefficient 2w  is obtained: 

 12
1,2

w
w


  (5) 

- Based on Eqs. (2) and (5), the value of the weight coefficient 3w  is obtained: 

 2 13
2,3 1,2 2,3

w w
w

  
 


 (6) 

- Based on Eq. (3), the value of the weight coefficient nw  is obtained: 

 1

1,2 2,3 1,...
n

n n

w
w

   


  
 (7) 

Step 5. Calculation of the value of the most significant criterion. If it is as follows: 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 7 

 

1

1

n

j

j

w



  (8) 

upon applying Eqs. (5-8), the result yields: 

 1 1 11
1,2 1,2 2,3 1,2 2,3 1,

... 1
... n n

w w w
w

      
    

   
 (9) 

From Eq. (9), the value of the weight coefficient of the most significant criterion is 

obtained as: 

 1

1,2 1,2 2,3 1,2 2,3 1,

1

1 1 1
1 ...

... n n

w

      



   
   

 (10) 

Step 6. Calculation of weight coefficients of other criteria. Applying Eqs. (5-7), the 

remaining weight coefficients of criteria, w2, w3,…,wn, are obtained.  

Step 7. Evaluation of quality of defined significance of the adjacent criteria. The 

relations of significance of the adjacent criteria need to be checked, to avoid as much as 

possible subjectivity of the decision makers. 

Step 7.1. Evaluation of quality of defined significance values. The evaluation of 

quality of defined significance values is made based on the relation of the significance of 

the most and the least significant criterion (η1,n). The value of the least significant 

criterion can be obtained from Eq. (4): 

 1

1,

k
n

n

w
w


  (11) 

where
k
nw  presents the control weight coefficient of the criterion Cn. 

The values wn and 
k
nw  should be approximately equal. If their deviation amounts to 

10% approximately, it can be concluded that the relations of significance of the adjacent 

criteria are defined with quality, and vice versa. Checking of the deviation is done by 

applying the expression: 

 1
n

n k
n

w
d

w
   (12) 

where dn presents the value of the deviation of the weight coefficients of the criterion Cn.  

If the condition 0 ≤ dn ≤ 0.1 is met, then the evaluations of the significance relations of 

the adjacent criteria are defined with quality, i.e. these meet the requirements. If dn > 0.1, 

it is necessary to define new relations between the criteria. However, since the research is 

usually an extensive process engaging significant resources, an additional step can be 

applied in order to find an error. In such cases, Step 7.2 is applied. 

Step 7.2. Additional evaluation of quality of the defined significance of the adjacent 

criteria. Before returning to defining relations of significance of the adjacent criteria, it is 



8 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

possible to make additional quality control. For this purpose, the step from Step 7.1 is 

repeated, in which the relations between the criteria Cn-1 and Cn-2 are defined.  

For this procedure, it is necessary for the decision-maker to define new relations: 

 11 1 1, 1 1, 1
1

: :1n n n
n

w
w w

w
   



    (13) 

 11 2 1, 2 1, 2
2

: :1n n n
n

w
w w

w
   



    (14) 

Then, it is calculated: 

 11
1, 1

k
n

n

w
w






  (15) 

 12
1, 2

k
n

n

w
w






  (16) 

where 1
k
nw  and 2

k
nw  are the control weight coefficients of the criterion Cn-1 respectively, 

Cn-2. 

Finally, the value is obtained: 

 
1

1

1

1 nn k
n

w
d

w






   (17) 

 
2

2

2

1 nn k
n

w
d

w






   (18) 

Here, dn-1 and dn-2 are the values of deviation of the weight coefficients of the criteria Cn-1 

and Cn-2, respectively.  

If the conditions  1 0, 0.1nd    and  2 0, 0.1nd    are met, it can be concluded that 
there was an error in defining the relations of significance between the most and the least 

significant criterion (η1,n). In that case, the existing results can be accepted or the values 

can be defined again (η1,n) and quality checked again as well. Briefly said, if 

 1 0, 0.1nd    or  2 0, 0.1nd   , the complete procedure of defining the relations of 
significance and calculation of the weight coefficients must be repeated.  

Step 8. Aggregation of weight coefficients of criteria for each expert. The previous 

seven steps refer to the definition of weight coefficients when decisions are made by a 

single decision maker, i.e. an expert. In situations where the decision is made by several 

people, it is necessary to aggregate their opinions. For the purposes of this model, the 

weigh coefficients of the criteria were aggregated, which were obtained for each expert 

separately, using the Bonferroni aggregator: 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 9 

 

, 1

1
( ) ( )

( 1)

l
a e p u q
i i i

e u
e u

w w w
l l




 
 

  
  

 

  (19) 

where l is the number of experts,
a
iw  are the aggregated values obtained by applying the 

Bonferroni aggregator, p,q ≥ 0 are the stabilization parameters of Bonferroni aggregator, 

e and u are the e-th or u-th expert, where 1 ≤ e, u ≤ l. 

2.2 Rough MABAC method 

The MABAC (Multi-Attributive Border Approximation area Comparison) method 

belongs to the group of newer and frequently applied methods. This method was 

developed by Pamučar i Ćirović [37]. In this paper, the method is improved by applying 

rough numbers. In the next part of the paper, the basic assumptions about rough numbers 

and the presentation of the steps of the MABAC method in rough environment are given. 

In the 1990s, Pawlak [38,39], developed rough sets. Some twenty years later, inspired 

by the idea of rough sets, Zhai et al. [40], developed rough numbers. Rough numbers 

prove to be very useful in considering uncertainty. A brief description of rough numbers 

is given below. 

Definition 1 [40]. Let’s assume that there is a set F consisting of (K1, K2, K3,…, Kt), 

which represents all the objects in a particular universe U. At the same time, there is Y 

presenting boundary objects of the universe U. If all the elements form part of the 

sequence K1<K2<K3<…<Kt, where it is valid that ,  ,  1qY U K F q t     , then the 

following parameters can be defined: 

  ( ) / ( )q qApr K Y U F Y K    (20) 

  ( ) / ( )q qApr K Y U F Y K    (21) 

 
 

   

( ) / ( )

/ ( ) / ( )

q q

q q

Bnd K Y U F Y K

Y U F Y K Y U F Y K

   

   
 (22) 

where ( )qApr K  is the lower approximation, and ( )qApr K  is the upper approximation, 

while Bnd(Kq) is boundary interval of the element Kq. 

Definition 2 [40]. The element Gq can be presented as a rough number (RN(Kq)). The 

rough number RN(Kq) is defined by its lower limit ( ( )qLim K ) and the upper limit 

( ( )qLim K ), where 

 q
1

( ) (Y) ( )q
L

Lim K F Y Apr K
N

   (23) 



10 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

 q
1

( ) (Y) ( )q
U

Lim K F Y Apr K
N

   (24) 

 ( ) ( ), ( )q q qRN K Lim K Lim K   
 (25) 

Here, NL and NU are the number of objects contained in ( )qApr K and ( )qApr K , 

respectively. Upper and lower limits indicate average value of the elements included into 

the upper and lower approximation, respectively. The difference between them represents 

the rough boundary interval (RBnd(Kq)): 

 q( ) ( ) ( )q qRBnd K Lim K Lim K   (26) 

Rough boundary interval represents uncertainty of the element Kq, where a higher 

number indicates higher imprecision, while a lower number indicates better precision, 

based on what subjective information only now can be marked with a rough number. 

By using the rough numbers shown, the classic MABAC method is modified. The 

steps of the rough MABAC method are given in the following [32-36]. 

Step 1. Forming of initial decision-making matrix (X). In the first step, the values of m 

alternatives are being defined by n criteria.  

 

1 2

11 12 11

221 222

31 323 3

1 2

( ) ( ) ( )

( )( ) ( )

( ) ( ) ( )

( ) ( ) ( )

n

n

n

n

m m m mn

C C C

RN x RN x RN xA

RN xRN x RN xA

RN x RN xX A RN x

A RN x RN x RN x

 
 
 
 
 
 
 
 

 (27) 

Step 2. Normalization of the elements of the initial decision-making matrix. In this 

step, the elements of the initial decision-making matrix (X) are normalized, in order to 

obtain the normalized matrix (N), by using the following expressions: 

 ,
ij j ij j

ij ij

j j j j

x x x x
t t

x x x x

 

   

 
 

 
 (28) 

 ,
ij j ij j

ij ij

j j j j

x x x x
t t

x x x x

 

   

 
 

 
 (29) 

By using Eq. (28) the normalization of the benefit type criteria is done, while using 

Eq. (29) the normalization of the cost type criteria is made, where: 

     
11

max , , min , for cost type criteriaj ij ij
i mi m

x x for benefit type criteria x


  
  (30) 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 11 

     
1 1
min , , max , for cost type criteriaj ij ij

i m i m
x x for benefit type criteria x



   
  (31) 

Step 3. The calculation of the elements of the weighted matrix (V) is performed by 

using the following equation: 

     1 , 1a aij i ij ij i ijv w x t v w x t     (32) 

Step 4. Defining border approximation area matrix (BAA). The BAA matrix (G) has a 

form n x 1, that is, the BAA s formed for each criterion separately, using the following 

equations: 

    
1/ 1/

1 1
,

m m
m m

j ij j ij
i i

g v g v
 

    (33) 

Step 5. Calculation of the matrix elements for alternatives distance from the BAA (Q) 

  ,ij ij
m x n

Q q q 
 

 (34) 

where: 

 
     

     

, ,
,

, ,

E ij j ij j

ij

E ij j ij j

d v g if RN v RN g
q

d v g if RN v RN g

 
 

  
   

 for benefit type criteria (35) 

 
     

     

, ,
,

, ,

E ij j ij j

ij

E ij j ij j

d v g if RN v RN g
q

d v g if RN v RN g

  
 

  
  

 for cost type criteria (36) 

  
   

   

2 2

2 2

,

,

, for cost type criteria

ij j ij j

E ij j

ij j ij j

v g v g for benefit type criteria

d v g

v g v g

 
   

 
  
 

    

 (37) 

Belonging of the alternative Ai to certain approximation area (G, G+ ili G-) is 

determined as follows: 

 

0

0

0

ij

i ij

ij

G if q

A G if q

G if q





 
 
 

  
 

  

 (38) 

Step 6. Ranking of alternatives. Final values of criteria functions are calculated by 

summing the elements of the matrix Q by rows: 



12 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

 

1 1

,  , 1, 2,..., , 1, 2,..., .

n n

i ij i ij

j j

s q s g j n i m

 

      (39) 

Step 7. To convert a rough number ( ) ( ), ( )i i iRN K Lim K Lim K   
into the crisp value, 

the following equation is used [35]: 

      min ( ) max ( ) min ( )crisp Ni i i i i
i ii

K Lim K K x Lim K Lim K   
  

 (40) 

For the needs of calculation of individual equation elements (40), the following is 

used [35]: 

 

 

   

 

   

( ) ( ), ( )

( ) min ( )
( )

max ( ) min ( )

( ) min ( )
( )

max ( ) min ( )

i i i

i i
i

i

i i
ii

i i
i

i

i i
ii

RN K Lim K Lim K

Lim K Lim K
Lim K

Lim K Lim K

Lim K Lim K
Lim K

Lim K Lim K

  
 

 
 
 
 
 

 
 

 
 

 (41) 

 
 ( ) 1 ( ) ( ) ( )

1 ( ) ( )

i i i iN
i

i i

Lim K x Lim K Lim K x Lim K
K

Lim K Lim K

 


 
 (42) 

3. CASE STUDY 

The decision supporting model, as shown above, is applied to the evaluation of the 

methods and tools of Lean concept of management. In the following two sections is given 

the description of the criteria and the calculation of the weight coefficients of the criteria, 

followed by a description of the alternatives and their ranking. 

3.1 Calculation of weight coefficients of criteria 

In the following, the weight coefficients of the criteria were obtained by applying the 

DIBR II method, considering that the DIBR II method is used for the first time in solving 

a real problem through group decision-making, a detailed description of all calculations is 

given below. 

Step 1. By analyzing available literature and with the help of four experts who deal 

with maintenance problems, five key criteria were reached according to which it is 

possible to evaluate the methods and tools of the Lean CoM management. These criteria 

are given in the following text. 

Reduction of a SPTS maintenance cycle time (C1), enables overview of basic factors 

in failures occurrence, distribution and duration of maintenance preparation and 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 13 

execution, and represents a basis for maintenance procedures establishment and analysis 

of overall efficiency and effectiveness of SPTS TM. 

Time preview of SPTS condition consist of their alternating states of operations and 

failures, which represent time intervals expressed in kilometers, motor hours or hours of 

SPTS work. Research in SPTS maintenance area is reduced to the analysis of causes of 

stated conditions. The analysis of those causes enables determination of basic parameter 

values, essential for designing the procedures and CoM of SPTS as a whole. 

Failure time consists of a number of time segments. The goal of all improvement 

procedures is to reduce time of waiting for the operation, by which maintenance cycle 

time is optimized.   

Optimization level of equipment and working positions arrangement (LAYOUT) 

within the maintenance system structure of technical systems for special purposes (C2). 

Proper and purposeful spatial arrangement of equipment and tools is very important for 

those who want to apply Lean management in organization systems, especially in 

production or SPTS maintenance. Layout helps the management in defining equipment, 

tools and work positions arrangement, in order to minimize circulation of SPTS, tools and  

equipment, at the expense of higher circulation of spare parts and  consumables, resulting 

in greater system operational readiness, due to shorter logistical and administrative 

waiting times, and lower transportation costs The assumption is that operators working in 

such a conceived organization system are qualified and trained to quickly and precisely 

perform all activities of the basic process of the system [8].  

SPTS effectiveness increase percentage, E(t) (C3) can be mathematically defined as 

probability that maintenance tasks will be performed in a certain period of time, with the 

prescribed functioning regime and with the implementation of prescribed maintenance 

activities - preventive/corrective maintenance programs, and it is expressed as the product 

of reliability P(t), functional suitability (suitability for functional use) FP(t) and 

availability (readiness) for intended use R(t). Each of the effectiveness components can 

take a value in the interval from 0 to 1. Therefore, the effectiveness can have the same 

values. The components of effectiveness and overall effectiveness of SPTS, should be 

projected and adopted already in the development phase, but their values should also be 

monitored and analyzed throughout the lifetime. Improvement is achieved if the 

effectiveness expressed in percentage is increased after the application of the alternatives. 

System efficiency increase level Efik (C4) can be defined as the ratio between the output 

and input, i.e. as the ratio between the products and resources (benefits and investments, 

needs and possibilities, planned and realized). It is also a number between 0 and 1.   

The mathematical interpretation of efficiency, in general, is as follows: 

 
 

 
fik

O output
E

I input
  (43) 

An improvement is achieved if the reached value is lower after the application of the 

alternatives, compared to the moment before the application of the alternatives, because 

the quantity being compared is the maintenance cycle time before and after the 

application of the alternatives.  

Reduction in number of executors (C5) is the goal of optimization of technological 

process, based on consideration of combining certain work positions, which are similar in 



14 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

terms of technological procedures, additional education and training of executors to 

perform more related operations by better planning, scheduling SPTS maintenance 

operations and eliminating unnecessary idling losses in TM of SPTS. 

Step 2. After defining the criteria, the experts performed criteria ranking by their 

significance. The criteria ranking by four experts Ee, where  1, 2, 3, 4e , is shown as 

follows:  

1 1 2 3 4 5

2 1 3 2 4 5

3 1 2 3 4 5

4 1 2 3 5 4

:

:

:

:

E C C C C C

E C C C C C

E C C C C C

E C C C C C

   

   

   

   

 

Step 3. Table 1 shows the assessment of the significance ratio performed by the 

experts. 

Table 1 Criteria significance ratios defined by experts 

Expert 1 Expert 2 Expert 3 Expert 4 

C1:C2=2.0 C1:C3=2.5 C1:C2=1.8 C1:C2=1.7 

C2:C3=1.1 C3:C2=1.2 C2:C3=1.0 C2:C3=1.3 

C3:C4=1.6 C2:C4=1.6 C3:C4=1.5 C3:C5=1.5 

C4:C5=1.3 C4:C5=1.4 C4:C5=1.1 C5:C4=1.1 

C1:C5=5.0 C1:C5=7.0 C1:C5=3.0 C1:C4=3.5 

 

Step 4. In this step, by using Eqs. (5-7), the relation of the most significant criterion to 

the other criteria is calculated. If the weight coefficient is marked with 
e
iw , where e  are 

the experts,  1, 2, 3, 4e , and i  is the criterion,  1, 2, 3, 4, 5i  , then the following 

relations are obtained: 

E1: 
1

1 1
2

2

w
w  ; 

1 1
1 1 1
3

2 1.1 2.2

w w
w  


; 

1 1
1 1 1
4

2 1.1 1.6 3.52

w w
w  

 
; 

1 1
1 1 1
5

2 1.1 1.6 1.3 4.576

w w
w  

  
 

E2:
2 2

2 1 1
2

2.5 1.2 3

w w
w  


; 

2
2 1
3

2.5

w
w  ; 

2 2
2 1 1
4

2.5 1.2 1.6 4.8

w w
w  

 
;  

2 2
2 1 1
5

2.5 1.2 1.6 1.4 6, 72

w w
w  

  
 

E3:  
3

3 1
2

1.8

w
w  ;   

3 3
3 1 1
3

1.8 1 1.8

w w
w  


;   

3 3
3 1 1
4

1.8 1 1.5 2.7

w w
w  

 
;   

3 3
3 1 1
5

1.8 1 1.5 1.1 2.97

w w
w  

  
 

E4: 
4

4 1
2

1.7

w
w  ; 

4 4
4 1 1
3

1.7 1.3 2.21

w w
w  


; 

4 4
4 1 1
4

1.7 1.3 1.5 1.1 3.647

w w
w  

  
; 

4 4
4 1 1
5

1.7 1.3 1.5 3.315

w w
w  

 
 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 15 

 

 

Step 5. By using Eq. (10), the weight coefficient of the most significant criterion is 

calculated. The detailed calculation for each expert is provided as follows: 

1
1

1
0.407

1 1 1 1
1

2 2.2 3.52 4.576

w  

   

 

2
1

1
0.478

1 1 1 1
1

3 2.5 4.8 6.72

w  

   

 

3
1

1
0.355

1 1 1 1
1

1.8 1.8 2.7 2.97

w  

   

 

4
1

1
0.382

1 1 1 1
1

1.7 2.21 3.647 3.315

w  

   

 

Step 6. Based on the ratios defined in the fourth step of this method, the weight 

coefficients of the other criteria were calculated: 

E1:
1
2

0.407
0.203

2
w   ; 

1
3

0.407
0.185

2.2
w   ; 

1
4

0.407
0.116

3.52
w   ; 

1
5

0.407
0.089

4.576
w    

E2: 
2
2

0.478
0.16

3
w   ;

2
3

0.478
0.191

2.5
w   ; 

2
4

0.478
0.1

4.8
w   ; 

2
5

0.478
0.071

6, 72
w    

E3: 
3
2

0.355
0.197

1.8
w   ; 

3
3

0.355
0.197

1.8
w   ; 

3
4

0.355
0.131

2.7
w   ; 

3
5

0.355
0.12

2.97
w    

E4: 
4
2

0.382
0.225

1.7
w   ; 

4
3

0.382
0.173

2.21
w   ; 

4
4

0.382
0.105

3.647
w   ; 

4
5

0.382
0.115

3.315
w    

Step 7. The weight coefficient of the least significant criterion was recalculated by 

applying the last (control) relation from Table 1 and the deviation from the obtained 

values was checked. The calculation was made using Eqs. (11) and (12). 

E1:  1 1 15 5 5
0.407 0.089

0.081 1 0.099 0, 0.1
5 0.081

w d d        ; 

E2:  2 2 25 5 5
0.478 0.071

0.068 1 0.044 0, 0.1
7 0.068

w d d        ; 

E3:  3 3 35 5 5
0.355 0.12

0.118 1 0.017 0, 0.1
3 0.118

w d d        ; 



16 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

E4:  4 4 45 5 5
0.382 0.115

0.109 1 0.055 0, 0.1
3.5 0.109

w d d        . 

Considering the fact that the provided deviation is less than 10%, it can be concluded that 

the experts were consistent in their opinions. 

Step 8. At the very end, the weight coefficients of the criteria were aggregated using 

standard Bonferroni aggregator and Eq. (19), with the value of the stabilization 

parameters p, q=1, which resulted in obtaining the final values of the weight coefficients 

of the criteria: 

1 2 3 4 50.405; 0.196; 0.187; 0.113; 0.099
a a a a a

w w w w w     . 

3.2 Ranking Alternative Solutions 

The methods and tools of Lean organization systems management, which also include 

the systems for TM of SPTS, represent the alternatives or possible solutions by 

implementation of which the improvement of the existing SPTS CoM condition is 

expected, reflected through the change (decrease or increase) of defined parameters 

(values) previously presented as criteria in the paper. By the help of experts, five 

alternative solutions were defined:  

Visual systems - VS (A1) present Lean method which with the help of visual aids 

(notice boards, and on warning lights about status of a process, displays, lines on the 

floors and walls of spare parts and consumables in the working environment) provides all 

the needed information for employees [5], related to: 

- basic system processes procedures; 

- costs; 

- required or achieved quality of the product/service; 

- product/service delivery/completion times; 

- current state of the process; 

- condition of tools; 

- number of employees; 

- workplace markings (yellow, green, red lines on plant or storage floors, etc.); 

- costs, attendance at work, injuries or safety at work and similar. 

By using this tool of Lean organization systems management, such environment is 

created where everything is precisely defined, from entry to exit, and where visual 

information is timely and useful [8]. 

Arrangement of workplace and surrounding area - 5 S (A2) is a set of rules for 

organizing a workplace for each employee. The name 5S was given by five English words 

starting with the letter “S”. The set of 5 S rules (Sort, Stabilize, Sustain, Shine, 

Standardize) refers to good maintenance of the plants in the organization system, offices, 

workplaces, spare parts and consumables warehouses, and it is useful for the analysis of 

improvements in an organization system [8]. The goal is for every work position inside 

organizational system to be arranged in such manner so it is maximally efficient, without 

unnecessary movement, making faster and easier work for each employee, by organizing 

all tools at their designed place, clearly visible, clean and ready for use in every moment. 

One of the most important 5 S method rule is that workers themselves take care of their 

workplace and thus contribute to the overall effectiveness of the process. Five S is the 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 17 

most recognizable method of the Lean concept [4] of managing organization systems 

because it is the easiest one to be applied and the results are immediately visible. By 

implementing the rules of the 5 S method, employees become encouraged in the easiest 

way to continue with the “lean“ transformation. Five S represents a method by which an 

excessive use of materials, energy, effort, etc. is reduced, while, on the other side, the 

quality and productivity of the system are raised to the optimal level. 

Computer maintenance management system – CMMS (A3) is reflected in the 

development, implementation and use of a software informational system that minimally 

does the following functions: management of work orders, maintenance planning, 

scheduling of maintenance activities, collection of data on maintenance history, 

management of approved funds in relation to costs, human resources management, 

management of spare parts and consumables for maintenance purposes, reporting on 

maintenance and consumption of spare parts and consumables, as well as on the level of 

their stock [8]. To achieve its purpose, a CMMS must be applied with complete and 

accurate data about SPTS, spare parts and consumables, plans, norms for maintenance 

activities and descriptions of maintenance activities, that is, all SPTS maintenance 

procedures.  

Work ticket system – KANBAN (A4) represents a system for making order in 

executions of process operations, control and management of spare parts stocks.  

The most common types of Kanban are [3]: 

- pull Kanban - circulates between spatially organized production structures and 

define the sequence of process activities between them; 

- production Kanban - circulates between workplaces within spatially organized 

production structures and define the quantities that must be produced in a given 

process, following the order of operations; 

- supplier Kanban –circulates between the producer and the supplier, and the delivery 

time to the producer, respectively, the consumer (customer), is pre-defined by 

instructions and procedures in the organization system. For example, commodity is 

transported once a day to the customer - every day at eight o'clock. The driver who 

hands over the commodity, that is, the realized Kanban for a certain delivery, takes 

over, at the same place, the Kanban for the next delivery. The quantity of goods, 

defined by the next Kanban, must be delivered to the customer the next day, at the 

same time.  

Kanbans can, with certain modifications, be used as warehouse record cards and 

identification cards in spare parts and consumables storages [3]. 

Continuous improvement – KAIZEN (A5) is the way towards continuous process 

improvements in order to eliminate losses.  

The word Kaizen originates from the Japanese language and it is composed of two 

Japanese words: kai - to separate and zen - to fix. The purpose of this Lean tool is 

recognition, separation, analysis and solution of problems, and then implementation and 

confirmation of that solution in practice. Kaizen bases its methods on the teachings of 

Edwards Deming and his PDCA (Plan-Do-Check-Act) quality circle. Small but constant 

improvements of the maintenance process are the goal of Kaizen tool implementation.  

Kaizen method involves all levels of employees, who are engaged in improvement 

process. Participants in Kaizen events are [8]: workers, heads of departments, mid-level 



18 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

managers, but also top managers. All of them, together in teamwork, should contribute to 

achieving gradual step by step improvements every day in every place. 

 The Kaizen approach encourages small daily continuous improvements and process 

that never ends, involving everyone’s progression, from workers to managers, using 

common sense as a basic principle for survival. Reasonable judgement and rational 

decision-making are essential for Kaizen approach.  

Balanced workforce engagement - NAGARA SYSTEM (A6) emphasizes human 

factor as very important in all organization systems, because there is no process without 

it, so it is essential to pay attention to effective and efficient use of human labor power. 

The Lean organization systems management resolves the effective and efficient use of 

human potential by applying the Nagara system tool [5], which requires 

“interdisciplinarity” from each individual (broader view of the work performed in the 

system, better understanding of work operations for which worker is not directly 

responsible), respectively, the ability to perform multiple tasks and serve multiple 

workplaces to a certain extent as needed, which contributes to increasing overall, not just 

work efficiency. Such practice reduces manpower to optimal without endangering 

working technology, work distribution is easier, and process flow more economical. The 

management of the organization system should take care of the timely personnel trainings 

for various related jobs in the organization systems in order to achieve their 

interdisciplinary skills. 

After defining the alternatives, the rough MABAC method is applied.  

Firstly, it is defined the initial decision-making matrix, as shown in the Table 2. 

Table 2 Initial decision-making matrix 

 C1 C2 C3 C4 C5 

A1 [6.2,6.5] S-M [13,15] H-VH [1,2] 

A2 [4.5,5] H-EH [10,12] S-M [3,4] 

A3 [6.8,7.2] ES-M [17,19] H-EH [2,3] 

A4 [7.3,7.5] H-VH [18,22] ES-M [1,2] 

A5 [3,3.7] M-H [12,16] M-H [2,3] 

A6 [3.2,4] VH-EH [15,20] VH-EH [4,5] 

 

As a part of the first step, the quantification of qualitative criteria was performed. It 

was done by using the scale shown in Table 3, consisting of seven linguistic descriptors. 

 

Table 3 Linguistic scale for quantifying of qualitative criteria 

Descriptor name Abbreviation Numeric value 

Extremely small ES 1 

Very small VS 2 

Small S 3 

Medium M 4 

High H 5 

Very high VH 6 

Extremely high EH 7 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 19 

By using the scale from Table 3, the initial decision-making matrix was quantified, as 

presented in Table 4.  

Table 4 Quantified initial decision-making matrix 

 C1 C2 C3 C4 C5 

A1 [6.2,6.5] [3,4] [13,15] [5,6] [1,2] 

A2 [4.5,5] [5,7] [10,12] [3,4] [3,4] 

A3 [6.8,7.2] [1,4] [17,19] [5,7] [2,3] 

A4 [7.3,7.5] [5,6] [18,22] [1,4] [1,2] 

A5 [3,3.7] [4,5] [12,16] [4,5] [2,3] 

A6 [3.2,4] [6,7] [15,20] [6,7] [4,5] 

By applying steps 2-7, the ranking of the alternatives is obtained. The final ranking of the 

alternatives is provided in Table 5. 

Table 5 Rank of alternatives  

 ( )iRN K  
crisp

iK  Rank 

A1 [0.15,0.478] 0.106 3 

A2 [0.028,0.406] 0.000 5 

A3 [0.226,0.647] 0.228 2 

A4 [0.317,0.705] 0.307 1 

A5 [-0.115,0.281] -0.143 6 

A6 [0.103,0.522] 0.096 4 

 

From Table 4, it is clearly visible that the alternative A4 is the first-ranked- the best, 

while the alternative A5 is the worst one.  

4. SENSITIVITY ANALYSIS 

Sensitivity analysis is an indispensable part of the entire decision-making process 

[41,42]. This segment of the decision-making model is mostly implemented through the 

changes of weight coefficients of criteria and can be found in a large number of recent 

researches such as [43-45]. The main purpose of sensitivity analysis is to determine how 

much the most influential criteria affect the final output - whether there are any and what 

the changes in ranking of alternatives are. In this particular case, the most influential is the 

criterion C1, which is two times more influential than the second-ranked C2. Considering 

significant difference between these two criteria, only changes to the criterion C1 were 

considered. For this purpose, 20 strategies with different weight coefficients of criteria 

were developed. The scenarios were created so that the value of the most significant 

criterion in each scenario was reduced by 5%, and the value that was reduced from the 

criterion C1 was evenly distributed among the other criteria. The values of the weight 

coefficients of the criteria according to the scenarios are given in Fig. 2. 



20 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

 

Fig. 2 Review of criteria weight coefficients changes  

According to the reviewed weight coefficients, the rough MABAC method was 

applied again and a new ranking was made for each scenario, as in Table 5. 

Table 5 Ranking alternatives using different scenarios 

  Si S1-S5 S6-S9 S10-S11 S12-S16 S17-S20 

A1 3 4 4 4 5 6 

A2 5 5 5 5 4 4 

A3 2 2 3 3 3 3 

A4 1 1 1 2 2 2 

A5 6 6 6 6 6 5 

A6 4 3 2 1 1 1 

 

Considering Table 5, it is easily noticeable that there are changes in the ranking of 

alternatives when the weight coefficients change. This clearly indicates that the model is 

sensitive to weight coefficients change. What is important is that it should not be overly 

sensitive, which is analyzed below:  

-  The alternative A1 is ranked from the fourth to sixth place. In relation to the initial 

ranking, the change already occurs in the first scenario, but it remains until the 

scenario S11. From the scenarios S12 to S20, this alternative moves to the fifth and 

sixth place. 

-  The alternatives A2, A3, A4 and A5 as a part of the scenario analysis have only one 

change in ranking. The alternative A2, from the scenario S12, moves from the fifth-

ranked place to the fourth-ranked alternative. The alternative A3 from the scenario 

S6, moves from the second-ranked place to the third-ranked alternative. The 

alternative A4 from the scenario S10 moves from the first-ranked position to the 



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 21 

second-ranked alternative. The alternative A5 from the scenario S17 moves from the 

sixth-ranked place to the fifth-ranked alternative. 

-  The alternative A6 has more significant changes. Namely, from the fourth-ranked 

alternative in the initial scenario (Si), it gradually takes the first place from scenario 

S10. 

-  As it can be observed, greater changes occur from the tenth scenario, when the most 

significant criterion is already reduced by 50%.  

-  The alternative A4, which was ranked first in the initial scenario, and due to major 

changes in the weight coefficients, retains the first ranking until the scenario S9, 

while from the scenario S10 until the end, it is the second-ranked. The lowest-ranked 

alternative, A6, retains this position in almost all scenarios. 

The above analysis clearly shows that there are changes in the ranking of alternatives 

following the changes in the weight coefficients of the criteria, but these changes are 

gradual. Accordingly, it is very clear that the presented model is quite stable. Regardless 

the stated, mathematical proof of the stability of the model was also conducted. By 

applying Spearman's rank correlation coefficient, it was defined whether the changes in 

the rankings of alternatives can be considered large and unacceptable. The equation for 

calculating Spearman's rank correlation coefficient is as follows:  

 

2

1

2

6

1
( 1)

n

i

i
rccS

n n




 



 (44) 

Here, Ξi is the difference of rank according to the given scenario and the rank in the 

corresponding scenario, and n is the number of ranked elements.  

The Spearman's rank correlation coefficient values are given in Table 6. 

Table 6 Spearman's coefficient values 

  Si S1-S5 S6-S9 S10-S11 S12-S16 S17-S20 

Si 1 0.943 0.886 0.657 0.543 0.371 

S1-S5   1 0.943 0.829 0.771 0.657 

S6-S9     1 0.943 0.886 0.771 

S10-S11       1 0.943 0.829 

S12-S16         1 0.943 

S17-S20           1 

 

The Spearman's coefficient of correlation of the ranking of considered strategies falls 

within the range of  0.371,1rccS  . The most important values of this coefficient are in 
relation of the initial strategy to other strategies. There it is observable that the Spearman 

coefficient is very high up to the strategy S11. Only from the strategy S12 to S16 it is 

significantly lower (0.513), but still quite high. From the strategy S17 up to S20, this 

value decreases more significantly, but still shows there is a certain correlation. This is 

expected since the most influential criterion is reduced to the minimum. Both 

mathematical and theoretical analysis clearly indicate that changes in alternatives rankings 



22 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

by different strategies, due to the changes in weight coefficients of the criteria, are gradual 

and expected.  Everything above mentioned confirms the conclusion of model stability. 

5. CONCLUSION 

This paper deals with the issue of ranking methods and techniques of Lean concept, 

with the purpose of improvement of work process and organization management of SPTS 

TM in the Army of Serbia.  Resulting conclusion suggests that in order to achieve the best 

results in conditions of limited opportunities for investment in the work process and 

organization of SPTS TM, should be used the best-ranked methods or techniques. 

For ranking alternative solutions, a hybrid model based on two methods, DIBR II and 

MABAC, was applied in a rough environment. Research showed that proposed model 

could successfully evaluate the methods and techniques of the Lean concept. Considering 

that the MABAC method was applied in a rough environment, the values from reality 

were very successfully reflected in the model.  

In this paper, the DIBR II method was applied for the first time in solving a specific 

case study. Expert evaluation of the criteria carried out in this research, showed that the 

way of comparing the criteria used in this method was very suitable for understanding and 

use by the experts who had never encountered this issue before. 

Presented model quality was tested in sensitivity analysis. Through this process, it was 

checked the way the changes in the weight coefficients affect the ranking of alternatives.  

It is clearly observable, through the analysis, that the model is sensitive enough, but not 

too sensitive. High correlation of alternatives rankings during different changes in weight 

coefficients of the criteria indicates the stability of the model. 

Basic limitation of the model is the fact that the experts were only able to use crisp 

values for comparison when defining the ratio of criteria.  In some situations, where 

experts were not certain, there was a clear space for the application of scientific fields that 

treat uncertainties well, such as fuzzy numbers, rough numbers, gray numbers, etc. The 

improvement of this DIBR II method should be the subject of future research. 

REFERENCES  

1. Epler, I.J., 2013, Models of technical system maintenance management, Military technical courier, 

61(1), pp. 178-195.  

2. Wireman, T., 2004, Benchmarking BEST Practices in Maintenance Management, Industrial Press, New 

York. 

3. Epler, I.J., Božičković, R.C., Arsić, S.N., Dinić, J.B., 2016, Real improvement processes in the army 

based on the lean six sigma concept, Military technical courier, 64(4), pp. 1132-1156.  

4. Božičković, R., 2005, Lean koncept u efektivnim proizvodnim sistemima, PhD Thesis, University of East 

Sarajevo, Faculty of Mechanical Engineering East Sarajevo, Bosnia and Herzegovina. 

5. Marić, B., 2010, Model upravljanja proizvodnim procesom u remontno – proizvodnim sistemima na 

bazi Lean koncepta, PhD Thesis, University of East Sarajevo, Faculty of Mechanical Engineering East 

Sarajevo, Bosnia and Herzegovina. 

6. Womack, J., Jones, D., Roos, D., 1991, The Machine that Change the World, Free Press, New York. 

7. Womack, J., Jones, D., 1996, Lean Thinking, Free Press, New York. 

8. Epler, I., 2017, Lean koncept u održavanju tehničkih sistema specijalne namene, PhD Thesis, University 

of East Sarajevo, Faculty of Mechanical Engineering East Sarajevo, Bosnia and Herzegovina.  



 Application of the DIBR II – rough MABAC Decision-Making Model for Ranking... 23 

9. Srinivasan, M.M., Bowers, M.R., Gilbert, K., 2014, Lean maintenance Repair on Overhaul, Mc Graw 

Hill Education.  

10. Tamanna, S., Rahman, M.M., 2022, Prioritization of Effective Lean Tools for Reliability Analysis & 

Maintenance Strategy, Journal of Engineering Advancements, 3(3), pp. 112-130. 

11. Singha Mahapatra, M., Shenoy, D., 2022, Lean maintenance index: a measure of leanness in 

maintenance organizations, Journal of Quality in Maintenance Engineering, 28(4), pp. 791-809. 

12. Bhebhe, M., Zincume. P.N., 2022, Application of Lean principles in railway maintenance, Proc. 33rd 

Annual Southern African Institute for Industrial Engineering Conference (SAIIE33), The Capital, 

Zimbali, South Africa, pp. 1134-1148. 

13. Korchagin, A., Deniskin, Y., Pocebneva, I., Vasilyeva, O., 2022, Lean Maintenance 4.0: implementation 

for aviation industry, Transportation Research Procedia, 63, pp. 1521-1533. 

14. Dragone, I.S., Biotto, C.N., Serra, S.M.B., 2021, Evaluation of Lean Principles in Building 

Maintenance Management, Proc. 29th Annual Conference of the International Group for Lean 

Construction (IGLC), 13-22, Lima, Peru.  

15. Puška, A., Maksimović, A., Stojanović, I., 2018, Improving organizational learning by sharing 

information through innovative supply chain in agro-food companies from Bosnia and Herzegovina, 

Operational Research in Engineering Sciences: Theory and Applications, 1(1), pp. 76-90. 

16. Riaz, M., Athar Farid, H.M., 2022, Picture fuzzy aggregation approach with application to third-party 

logistic provider selection process, Reports in Mechanical Engineering, 3(1), pp. 227–236. 

17. Bošković, S., Švadlenka, L., Dobrodolac, M., Jovčić, S., Zanne, M., 2023, An Extended AROMAN 

Method for Cargo Bike Delivery Concept Selection, Decision Making Advances, 1(1), pp. 1–9. 

18. Ali, A., Ullah, K., Hussain, A., 2023, An approach to multi-attribute decision-making based on 

intuitionistic fuzzy soft information and Aczel-Alsina operational laws, Journal of Decision Analytics 

and Intelligent Computing, 3(1), pp. 80–89. 

19. Mahmood, T., ur Rehman, U., 2023, Bipolar complex fuzzy subalgebras and ideals of BCK/BCI-

algebras, Journal of Decision Analytics and Intelligent Computing, 3(1), pp. 47–61. 

20. Ashraf, A., Ullah, K., Hussain, A., Bari, M., 2022, Interval-Valued Picture Fuzzy Maclaurin Symmetric 

Mean Operator with application in Multiple Attribute Decision-Making, Reports in Mechanical 

Engineering, 3(1), pp. 210–226. 

21. Maurya, J.K., Mishra, S.K., 2022, Strong Complementary Approximate Karush-Kuhn-Tucker 

Conditions for Multiobjective Optimization Problems, Yugoslav Journal of Operations Research, 32(2), 

pp. 219-234. 

22. Badi, I., Stević, Ž., Bayane Bouraima, M., 2023, Overcoming Obstacles to Renewable Energy 

Development in Libya: An MCDM Approach towards Effective Strategy Formulation. Decision Making 

Advances, 1(1), pp. 17–24. 

23. Božanić, D., Pamucar, D., 2023, Overview of the Method Defining Interrelationships Between Ranked 

Criteria II and Its Application in Multi-criteria Decision-Making, In: Chatterjee, P., Pamucar, D., 

Yazdani, M., Panchal, D. (Eds.) Computational Intelligence for Engineering and Management 

Applications. Lecture Notes in Electrical Engineering, 984, Springer, Singapore. pp. 863-873. 

24. Badi, I., Abdulshahed, A., 2021, Sustainability performance measurement for Libyan Iron and Steel 

Company using Rough AHP, Journal of Decision Analytics and Intelligent Computing, 1(1), pp. 22–34.  

25. Roy, J., Adhikary, K., Kar, S., Pamucar, D., 2018, A rough strength relational DEMATEL model for 

analysing the key success factors of hospital service quality, Decision Making: Applications in 

Management and Engineering, 1(1), pp. 121-142. 

26. Qi, J., Hu, J., Peng, J.-H., 2020, Integrated rough VIKOR for customer-involved design concept 

evaluation combining with customers’ preferences and designers’ perceptions, Advanced Engineering 

Informatics, 46, 101138. 

27. Song, W., Ming, X., Wu, Y., Zhu, B., 2014, A rough TOPSIS Approach for Failure Mode and Effects 

Analysis in Uncertain Environments, Quality and Reliability Engineering, 30(4), pp. 473-486. 

28. Arsić, S.N., Pamučar, D., Suknović, M., Janošević, M., 2019, Menu evaluation based on rough 

MAIRCA and BW methods, Serbian Journal of Management, 14(1), pp. 27-48.  

29. Jangid, V., Kumar, G.A., 2022, Novel Technique for Solving Two-Person Zero-Sum Matrix Games in a 

Rough Fuzzy Environment, Yugoslav Journal of Operations Research, 32(2), pp. 251-278. 

30. Pamucar, D., Žižović, M., Đuričić, D., 2022, Modification of the CRITIC method using fuzzy rough 

numbers, Decision Making: Applications in Management and Engineering, 5(2), pp. 362–371. 



24 D. BOŽANIĆ, I. EPLER, A. PUŠKA, S. BISWAS, D. MARINKOVIĆ, S. KOPRIVICA 

31. Sahu, R., Dash, S.R., Das, S., 2021, Career selection of students using hybridized distance measure 

based on picture fuzzy set and rough set theory, Decision Making: Applications in Management and 

Engineering, 4(1), 104–126. 

32. Chakraborty, S., Dandge, S.S., Agarwal, S., 2020, Non-traditional machining processes selection and 

evaluation: A rough multi-attributive border approximation area comparison approach, Computers & 

Industrial Engineering, 139, 106201. 

33. Tešić, D., Radovanović, M., Božanić, D., Pamucar, D., Milić, A., Puška, A., 2022, Modification of the 

DIBR and MABAC Methods by Applying Rough Numbers and Its Application in Making Decisions, 

Information, 13(8), p. 353. 

34. Tešić, D.Z., Božanić, D.I., Miljković, B.D., 2023, Application of MCDM DIBR-Rough MABAC Model 

for Selection of Drone for Use in Natural Disaster Caused by Flood. In: Filipovic, N. (eds) Applied 

Artificial Intelligence: Medicine, Biology, Chemistry, Financial, Games, Engineering. AAI 2022. 

Lecture Notes in Networks and Systems, 659, Springer, Cham, pp. 151-169. 

35. Roy, J., Chatterjee, K., Bandhopadhyay, A., Kar, S. (2018). Evaluation and selection of Medical 

Tourism sites: A rough Analytic Hierarchy Process based Multi Attributive Border Approximation area 

Comparison approach, Expert Systems, 35(1), e12232. 

36. Chattopadhyay, R., Das, P.P., Chakraborty, S., 2022, Development of a Rough-MABAC-DoE-based 

Metamodel for Supplier Selection in an Iron and Steel Industry, Operational Research in Engineering 

Sciences: Theory and Applications, 5(1), pp. 20–40. 

37. Pamučar, D., Ćirović, G., 2015, The selection of transport and handling resources in logistics centers 

using Multi-Attributive Border Approximation area Comparison (MABAC), Expert Systems with 

Applications, 42(6), pp. 3016-3028. 

38. Pawlak, Z., 1991, Rough sets-theoretical aspects of reasoning about data; Kluwer Academic Publishers, 

Dordrecht, Netherlands. 

39. Pawlak, Z., 1996, Rough sets, rough relations and rough functions, Fundamenta Informaticae, 27(2-3), 

pp. 103-108. 

40. Zhai, L.-Y., Khoo, L.-P., Zhong, Z.-W., 2009, A rough set based QFD approach to the management of 

imprecise design information in product development, Advanced Engineering Informatics, 23(2), pp. 

222–228. 

41. Čubranić-Dobrodolac, M., Švadlenka, L., Čičević, S., Trifunović, A., Dobrodolac, M., 2020, Using the 

Interval Type-2 Fuzzy Inference Systems to Compare the Impact of Speed and Space Perception on the 

Occurrence of Road Traffic Accidents, Mathematics, 8(9), 1548. 

42. Tešić, D.Z., Božanić, D.I., Pamučar, D.S., Din, J., 2022, DIBR - Fuzzy MARCOS model for selecting a 

location for a heavy mechanized bridge, Military technical courier, 70(2), pp. 314-339. 

43. Ulutaş, A., Topal, A., Pamučar, D., Stević, Ž., Karabašević, D., Popović, G., 2022, A New Integrated 

Multi-Criteria Decision-Making Model for Sustainable Supplier Selection Based on a Novel Grey WISP 

and Grey BWM Methods, Sustainability, 14(24), 16921. 

44. Radovanović, M., Petrovski, A., Žnidaršič, V., Ranđelović, A., 2021, Application of the Fuzzy AHP -

VIKOR Hybrid Model in the Selection of an Unmanned Aircraft for the Needs of Tactical Units of the 

Armed Forces, Scientific Technical Review, 71(2), pp. 26-35. 

45. Radovanović, M., Milić, A., Petrovski, A., 2021, Analysis of accuracy and precision of shooting with 

home: Made automatic rifles using the AHP method, Scientific Technical Review, 71(1), pp. 30-37. 

https://oresta.org/