Application of Fuzzy TOPSIS Method For Selection The Additional Protecting System On Railway Crossing

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/328281848

Application of fuzzy TOPSIS method for selection the additional protecting


system on railway crossing

Conference Paper · October 2018

CITATIONS READS

0 31

4 authors, including:

Milan Milosavljević
Academy of Technical and Art Applied Studies Serbia, Belgrade
19 PUBLICATIONS   41 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Milan Milosavljević on 15 October 2018.

The user has requested enhancement of the downloaded file.


APPLICATION OF FUZZY TOPSIS METHOD FOR SELECTION
THE ADDITIONAL PROTECTING SYSTEM ON RAILWAY
CROSSING
Milan MILOSAVLJEVIĆ1
Sandra KASALICA2
Dušan JEREMIĆ3
Goran VIL4

Abstract – Railway crossings, as places of crossroads of rail and road traffic, represent the sites of
potential conflicts, and as such, require continuous monitoring and implementation of adequate
measures that lead to increased safety due to elimination of accidents. For that reason, railway
crossings need to be properly marked and protected with appropriate protection system. The goal
of this paper is to improve the railway of safety at the railway crossing with active protection, by
selection additional insurance system. This paper presents the application of fuzzy TOPSIS method
on railway crossing PBK-2 on railway line Belgrade - Šid - state border. This approach helps in
solving problems where the vagueness and subjectivity are handled with linguistic values
parameterized by triangular fuzzy numbers. To validate the results, the sensitivity analysis was
performed through changing the weight coefficients of the criteria. The results of the sensitivity
analysis show that the proposed method gives valid and stable results.

Keywords – railway crossing, fuzzy, TOPSIS.

1. INTRODUCTION activated when the train comes in. These are light
signals, bumpers and half-barriers. Active traffic
Railway crossings are places of conflict where
control devices are also supplemented with signals
traffic accidents often occur. They are weak point in
used in passive traffic control.
road and railway infrastructure that seriosly
This paper combines TOPSIS method and fuzzy
jeopardizing traffic safety. A lot of accidents in traffic
logic, for creation hybrid method which is used in [2],
with serious consiquences, material damage and
[3] and [4]. The authors of [5] presented the ANFIS
fatalities are occur on railway crossings [1].
model, which supports the process of selecting which
The main purpose of the traffic control device is to
railway crossings should receive an investment of
provide appropriate warning to drivers using
safety equipment and which was tested on 88 level
appropriate visual and audio devices and to help them
crossings in Serbia. In paper [6] authors presented
in undertaking correct activities to avoid accidents at
how and to which extent certain influential parameters
railway crossings. Traffic control devices are divided
cause accident mechanisms on railway crossings,
into active and passive. The basic conventional
some of which are used in this paper.
warning devices in Serbia are the Andreas cross, the
stop sign (passive) and half-barriers with light
2. FUZZY TOPSIS METHOD
signaling (active) [1].
In the literature we can find various additional The TOPSIS method is a multiple criteria decision
passive traffic control measures. Some of them are: making technique proposed by [7] to identify a
marking a roadway in front of a railway crossing, solution from a finite set of options. The fuzzy
enhanced warning signs, speed bumps and ramps in TOPSIS technique was proposed by [8] and this
front of the crossing in order to reduce the speed of method has nine important steps:
the road vehicle and other measures. Step 1: Form a committee of decision-makers, then
Active traffic control devices are devices that are identify the evaluation criteria.
1
High railway school of vocational studies, Zdravka Čelara 14, Belgrade, Serbia, mimilan89@gmail.com.
2
High railway school of vocational studies, Zdravka Čelara 14, Belgrade, Serbia, sandra.kasalica@gmail.com
3
High railway school of vocational studies, Zdravka Čelara 14, Belgrade, Serbia, jeremicd@gmail.com
4
JKP Gradsko saobraćajno preduzeće „Beograd“, Knjeginje Ljubice 29, Belgrade, Serbia, goran.vil@gsp.co.rs
57
RAILCON ’18 APPLICATION OF FUZZY TOPSIS METHOD FOR SELECTION THE …
Step 2: Choose the appropriate linguistic variables Step 7: Calculate the distance of each alternative
for the importance weight of the criteria and the from FPIS and FNIS, respectively. The separation of
linguistic ratings for alternatives with respect to each alternative from the A*, A- is given as:
criteria.
( )
n
d i =  d v~ij , v~j * , i = 1,2,..., m
Step 3: Aggregate the weight of criteria to get the *
(12)
aggregated fuzzy weight w ~ of criterion Cj, and pool j =1
j

( )
the decision makers’ opinions to get the aggregated n
d i =  d v~ij , v~j − , i = 1,2,..., m

fuzzy rating ~
xij of alternative Ai under criterion Cj. (13)
j =1

Rk = (ak , bk , ck ), k = 1,2,3,..., K
~
(1) where d(.;.) represents distance between two fuzzy
numbers.
than aggregate fuzzy ranking can be calculated The distance measurement between two fuzzy
using equations: numbers can be calculated as [1-8]:
1 K
a = min k (ak ), b =  bk , c = max k (ck ) (2) ( ) ~
d a~, b =
1

(a1 − b1 )2 + (a2 − b2 )2 + (a3 − b3 )2  (14)
k k =1 3
Step 4: Construct the fuzzy decision matrix and the Step 8: Calculate the closeness coefficient of each
normalized fuzzy decision matrix. alternative. A closeness coefficient is defined to
~
R= ~
rij mxn
, i = 1,2,3,..., m; j = 1,2,3,..., n (3)
determine the ranking order of all alternatives once
* −
the di and d i of each alternative Ai(i=1,2,…,m) has
where B and C are the set of benefit criteria and been calculated. The closeness coefficient of each
cost criteria, respectively, and alternative is calculated as:

a b c  di
rij =  ij* , ij* , ij* , j  B
~ (4) CCi = −
, i = 1,2,...m (15)
c c c  di + di
*
 j j j 

Step 9: According to the closeness coefficient, the


 a j− a j− a j− 
rij =  , j  C
~ ranking order of all alternatives can be determined.
, , (5)
 cij bij aij  The best alternative has a highest value.
 
3. APPLICATION OF FUZZY TOPSIS
c j = max i cij , j  B
*
(6)
METHOD

a j = mini aij , j  C (7) The proposed model was used to determine the
additional protecting system, which should be placed
Step 5: Construct the weighted normalized fuzzy on the railway crossing PBK-2. The observed crossing
decision matrix. Considering the different importance is located on the open track and it is equipped with an
of each criterion, we can construct the weighted automatic device for securing traffic at railway
normalized fuzzy decision matrix as: crossings, as well as half-barriers with electrically

 
~ powered devices and light path signals.
V = v~ij mxn
, i = 1,2,3,..., m; j = 1,2,3,..., n (8) In order to increase the level of security, in
addition to the existing insurance system, three
v~ij = ~
rij W (9) experts proposed four types of additional protecting
system for the road crossing [10]:
Step 6: Determine the fuzzy positive-ideal (FPIS) 1. video surveillance (A1) - the camera is
and fuzzy negative-ideal (FNIS) solutions [9]. installed at the railway crossing, and serves to
(
A* = v~1* , v~2* ,..., v~n* = ) detect irregular behavior in traffic and reduce
accident.
(max v i  B ), (min v i  C )
(10)
ij ij
2. Separators (A2 and A3) - Separators are
physical barriers that are placed in the middle
A = (v~ , v~ ,..., v~ ) =

1− 2− n−
of the roadway on both sides of the railway

(min v i  B ), (max v i  C )
(11) crossing, as a supplement to the system with
ij ij half-barriers, in order to prevent the going
around of half-barriers. In this paper two
where B is associated with benefit criteria, and C is types of separators, metal (A2) and plastic
associated with cost criteria. (A3) are observed.

58
XVIII International Scientific-expert Conference on Railways Serbia, Niš, October 11-12, 2018
3. rumble strips (A4) - used as an additional Tab. 1. Linguistic variables for the ratings
protecting and warning device to alert drivers
Lingvistic terms Fuzzy numbers
of road vehicle to emerging potential hazards. Very low (VL) (1,1,2)
Defining the criteria for evaluating alternatives is Low (L) (1,2,3)
one of the most important decision-making tasks. The Medium (M) (2,3,4)
most important criteria are: High (H) (3,4,5)
1. possibility of controlling the behavior of Very high (VH) (4,5,5)
participants in road traffic,
2. technical possibility of installing the system, Based on first three steps from fuzzy TOPSIS
3. the cost of installing the system, algorithm mentiond above, linguistic values of the
4. the possibility of restricting participants in importance weight of the criteria and the linguistic
road traffic to irregular behavior, ratings for alternatives with respect to criteria are
5. the estimated lifetime of the technical devices shown in tables 2 and 3.
after which there are increased defects that Tab. 2. The importance weight of the criteria
leads to the replacement of the existing
device, D1 D2 D3
6. the influence of the technical solution on the C1 M (2,3,4) L (1,2,3) L (1,2,3)
psychological moment in the participants in C2 H (3,4,5) M (2,3,4) M (2,3,4)
road traffic. C3 H (3,4,5) VH (4,5,5) H (3,4,5)
Considering that all the criteria and their weights
C4 VH (4,5,5) M (2,3,4) H (3,4,5)
are descriptive (linguistic), they can be represented by
fuzzy linguistic descriptors, which is presented in C5 H (3,4,5) M (2,3,4) L (1,2,3)
table 1. C6 H (3,4,5) VH (4,5,5) L (1,2,3)
Tab. 3. The ratings of three candidates by decision makers under all criteria
Decision
Alternatives/Criteria C1 C2 C3 C4 C5 C6
makers
A1 L (1,2,3) M (2,3,4) VH (4,5,5) M (2,3,4) L (1,2,3) H (3,4,5)
A2 H (3,4,5) H (3,4,5) M (2,3,4) VH (4,5,5) H (3,4,5) M (2,3,4)
D1
A3 M (2,3,4) VH (4,5,5) M (2,3,4) H (3,4,5) VH (4,5,5) VL (1,1,2)
A4 L (1,2,3) VH (4,5,5) L (1,2,3) VL (1,1,2) M (2,3,4) VL (1,1,2)

A1 M (2,3,4) H (3,4,5) H (3,4,5) L (1,2,3) M (2,3,4) VH (4,5,5)


A2 H (3,4,5) M (2,3,4) M (2,3,4) VH (4,5,5) M (2,3,4) H (3,4,5)
D2
A3 L (1,2,3) H (3,4,5) L (1,2,3) VH (4,5,5) H (3,4,5) L (1,2,3)
A4 VL (1,1,2) VH (4,5,5) VL (1,1,2) L (1,2,3) L (1,2,3) L (1,2,3)

A1 L (1,2,3) M (2,3,4) VH (4,5,5) M (2,3,4) M (2,3,4) VH (4,5,5)


A2 M (2,3,4) M (2,3,4) H (3,4,5) H (3,4,5) H (3,4,5) M (2,3,4)
D3
A3 H (3,4,5) H (3,4,5) M (2,3,4) H (3,4,5) M (2,3,4) M (2,3,4)
A4 L (1,2,3) VH (4,5,5) L (1,2,3) L (1,2,3) M (2,3,4) VL (1,1,2)
and final ranking of alternatives are shown in table 4.
Next step is construction the weighted normalized
fuzzy decision matrix and calculation the (FPIS) and Tab. 4. Results of fuzzy TOPSIS model
(FNIS):
Di* Di- CCi Rank
(0,4;1,7;4,0 ), (1,2;2,9;5,0 ),  A1 2.758 5.915 0.682008 2
A* = (0,6;0,9;1,7 ), (1,2;3,7;5,0 ),  (16) A2 1.667 6.931 0.806098 1
(0,4;2,4;5,0 ), (0,6;3,4;5,0 ) A3 3.624 5.315 0.594551 3
A4 8.218 0.128 0.015373 4
(0,4;1,7;4,0 ), (1,2;2,9;5,0 ),  As we can see from table 4, the best alternative is
A = (0,6;0,9;1,7 ), (1,2;3,7;5,0 ), 
*
(17) A2 metal separator, and the worst is alternative A4.
(0,4;2,4;5,0 ), (0,6;3,4;5,0 ) 4. SENSITIVITY ANALYSIS
The distance of each alternative from and Di* Di- Special attention in this paper is committed to
can be calculated by eqations (12-14). Calculations sensitivity analysis, which can determine the stability

59
RAILCON ’18 APPLICATION OF FUZZY TOPSIS METHOD FOR SELECTION THE …
of the observed solution. The sensitivity analysis was of the criteria. Table 5 gives the scenarios of changing
done by changing the values of the weight coefficients the weight coefficients of the criteria.
Tab. 5. Scenarios with different criteria weights
Criterion Weight
Scenarios
C1 C2 C3 C4 C5 C6
S1 Uniform Weight Criteria 0.1667 0.1667 0.1667 0.1667 0.1667 0.1667
S2 Priority of Criterion C1 0.5 0.1 0.1 0.1 0.1 0.1
S3 Priority of Criterion C2 0.1 0.5 0.1 0.1 0.1 0.1
S4 Priority of Criterion C3 0.1 0.1 0.5 0.1 0.1 0.1
S5 Priority of Criterion C4 0.1 0.1 0.1 0.5 0.1 0.1
S6 Priority of Criterion C5 0.1 0.1 0.1 0.1 0.5 0.1
S7 Priority of Criterion C6 0.1 0.1 0.1 0.1 0.1 0.5
Combined Priority of
S8 0.1 0.2 0.3 0.2 0.1 0.1
Criterion C2, C3 and C4
The ranking of the alternative for different Based on the table 6 we can see that alternative A2
scenarios are presented in table 6, where S0 is the rank is the best solution in six of nine cases (67%) with
of alternative obtained using the actual weight of the values in the range of 0,66-0,86, which is an adequate
criterion. By analyzing the obtained results, it can be verification of the proposed model. Alternative A4 in
concluded that there is significant stability in most cases. all models are also on the same position, the last place.
Tab. 6. Alternatives ranking for different weight criteria scenarios
Alternatives S0 S1 S2 S3 S4 S5 S6 S7 S8
A1 2 2 3 3 1 3 3 1 2
A2 1 1 1 2 2 1 1 2 1
A3 3 3 2 1 3 2 2 3 3
A4 4 4 4 4 4 4 4 4 4
[3] Nadaban, S., Dzitac, S., & Dzitac, I., Fuzzy TOPSIS: A
5. CONCLUSION general view, Information Technology and
The obtained results using the fuzzy TOPSIS Quantitative Management, 91, 823-831, 2016.
method show that the presented method can be used to [4] Sun, C. C., A performance evaluation model by
support decision makers when choosing a system of integrating fuzzy AHP and fuzzy TOPSIS methods,
Expert systems with applications, 37(12), 7745-7754, 2010.
additional protecting system. Through the sensitivity
[5] Ćirović, G., Pamučar, D., Decision support model for
analysis, it has been shown that the solution has prioritizing railway level crossings for safety
stability, and through which the subjectivity and improvements: Application of the adaptive neuro-fuzzy
imprecision that are occurring every day in decision system. Expert Systems with Applications, 40, 6, 2208-
making are reduced. In addition to the example 2223, 2013.
shown, the advantage of this model is its ability to [6] Starčević, M., Barić, D., & Pliko, H., Survey-based
solve other decision-making problems. impact of influencing parameters on level crossings
The future development of the model can be seen safety, Promet - Traffic & Transportation, 28(6), 639-
in a more detailed analysis of the input parameters and 649, 2016.
[7] Hwang, C. L., & Yoon, K., Lecture Notes in
the formation of larger groups of experts that would
Economics and Mathematical Systems: Multiple
improve the input values. In the future period, Attribute Decision Making: Methods and Application,
presented model will be tested using another methods Springer Verlag, 1981.
of multi criteria decision making combined with fuzzy [8] Chen, C. T., Extensions of the TOPSIS for group
logic or rough numbers theory. decision-making under fuzzy environment, Fuzzy sets
and systems, 114(1), 1-9, 2000.
REFERENCES [9] Yu, X., Guo, S., Guo, J., & Huang, X., Rank B2C e-
commerce websites in e-alliance based on AHP and
[1] Kasalica, S., Safety improvement of traffic at railway
fuzzy TOPSIS, Expert Systems with Applications,
crossings. PhD thesis. University of Belgrade, Faculty
38(4), 3550-3557, 2011.
of mechanical engineering, 2013.
[10] Vil, G., Analiza stepena bezbednosti saobraćaja na
[2] Kang, D., Jang, W., & Park, Y., Evaluation of e-
putnom prelazu PBK-2. Specijalistički rad, Visoka
commerce websites using hierarchical TOPSIS based
železnička škola strukovnih studija, Beograd, 2016.
on E-S-QUAL, Applied Soft Computing, 42, 53-65, 2016.

60

View publication stats

You might also like