Using a New Fuzzy MADM Method to Improve the Number of

Vertical Handovers in the Case of Wireless Technology Selection for

VANETs

Abdeslam Houari, Tomader Mazri

Laboratory of Advanced Systems Engineering, Faculty of sciences, Ibn Tofail science University

Kenitra, Morocco

Keywords: VANET, ITS,5G, DSRC, MADM, Fuzzy, AHP, FAHP, TOPSIS.

Abstract: VANET is a promising project in the transportation field, and more precisely, in the intelligent transportation

(ITS) area. Its heterogeneous architecture has led researchers to use vertical handover to allow vehicles to

switch from one wireless technology to another (such as 5G, DSRC…) at any time and in any situation without

losing connection. For this purpose, several methods have been developed, among them the Multi-attribute

decision-making (MADM) methods, which allow the enhancement of decision-making in the vertical

handover process. This paper proposes a new approach for wireless technology selection based on an

improved TOPSIS method applied to order the alternatives. Simulation experiments have been conducted to

evaluate our approach, and the results show that our TOPSIS* method is more efficient than the classical

Fuzzy TOPSIS.

1 INTRODUCTION

Nowadays, the increasing number of accidents and

traffic jams on our roads have motivated the

automotive industry to increase the autonomy of

vehicles, make the vehicle's path as safe as possible,

and protect human life. For this reason, researchers in

the automotive field have turned to intelligent

transport systems (ITS). VANET (vehicular ad-hoc

network) network is a specific case of MANET

(Mobile Ad-hoc Networks) networks, which

researchers consider a promising project in the ITS

field. The idea is to interconnect vehicles and share

resources and information between them to explore

their surroundings better and cope with the different

threats and issues that may arise on the way. In a

VANET environment, vehicles communicate with

each other via V2V(vehicle-to-vehicle) mode and

with the infrastructure via V2I (vehicle-to-

infrastructure) mode, which enables them to

exchange with RSUs (Road-Side-Unit) and base

stations to take advantage of several services such as

internet access [1]. However, despite the multitude

of technologies available (5G, 6G, DSRC ...),

VANET faces several challenges; one of the most

critical is the connection loss [2], caused by the high

speed of the vehicles and the dynamic topology of the

network. In order to overcome this issue, researchers

have been interested in the vertical handover,

allowing to pass from one technology (support) to

another without loss of connection. This operation is

focused on the selection of the best technology in a

heterogeneous system such as VANET. For this

purpose, we opt-in this article for the MADM (Multi-

attribute decision-making) algorithms which have

proven their efficiency in several fields. MADM is

applied to select the best possible choice during the

vertical handover phase by considering various

decision criteria. In the MADM approach, there are

several algorithms such as DIA (Distance to Ideal

Alternative), ANP, and AHP/FAHP (Fuzzy Analytic

Hierarchy Process) for measuring the criteria weight,

and other algorithms for ranking the alternatives

(networks and technologies) such as VIKOR,

TOPSIS, and GRA. In this paper, we propose an

improvement of the TOPSIS algorithm (TOPSIS*) to

classify the available technologies to reduce the

number of vertical handovers and improve the quality

of service (QoS).

Houari, A. and Mazri, T.

Using a New Fuzzy MADM Method to Improve the Number of Vertical Handovers in the Case of Wireless Technology Selection for VANETs.

DOI: 10.5220/0010735000003101

In Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning (BML 2021), pages 387-391

ISBN: 978-989-758-559-3

387

2 RELATED WORK

Technology (support) selection is a fundamental step

that requires a dynamic selection of the best support at

a given time depending on the situation in which a

vehicle is confronted because mobility in a VANET

context directly impacts the topology and performance

of the used protocols [3]. MADM has been presented

as the most promising method to solve the alternative

selection problem. It is easy to implement using

simple mathematic methods and does not require any

specific physical resources.

In addition, MADM methods are widely used for

decision-making in the context of VHO; the most

important methods are SAW, TOPSIS, and VIKOR

[2] [4]. The main principle is to rank the alternatives

according to the score of the measured weights. Some

authors, such as [5], proposed arbitrary weights to

identify the importance of each attribute (criterion) by

QoS class. Others [6] were able to use the AHP to

calculate the weight of the criterion vectors and apply

the TOPSIS method to rank the alternatives (LTE, 4G,

5G ....). The results showed that the weight of the

criterion vectors is important in the decision-making

process.

Nevertheless, intelligent computing algorithms are

still the most efficient since they use intelligent

implementation techniques such as Fuzzy Logic and

neural networks. Fuzzy logic is useful for VH

decision-making because it can deal with radio signal

inaccuracy, user preferences, and QoS parameters.

Several authors [7] [8] have studied the Fuzzy AHP

and Fuzzy TOPSIS combination to measure the

relative weights of the evaluation criteria and classify

the alternatives as an improved solution to a problem

of inter-vehicle communications. This is why we have

chosen the fuzzy approach in our technology selection

model.

3 SYSTEM MODEL

In this study, we attempted to overcome the

weaknesses of the FTOPSIS method for application

fields characterized by high mobility, as is the case of

VANET. One of the major concerns this method faces

is the reversal phenomenon, which occurs at the

preference order level due to the addition or removal

of an alternative from the original decision problem.

The authors have made several attempts [9] to

improve the TOPSIS method, but no effective

solution is implemented yet.

The method we propose (TOPSIS*) is improving

the fuzzy TOPSIS method based on the vertical

handover decision by combining it with the Fuzzy

AHP to generate the criteria weights.

As shown in Figure 1, we collect the evaluation

criteria and the alternatives chosen for this study;

then, we build the decision matrix using the

information recovered from the first step. Once this is

done, the pairwise comparison process is initiated for

each QoS class. In this study, we considered the

following evaluation criteria: data rate, latency,

throughput, and coverage which will be processed as

weight vectors by the Fuzzy AHP method, and

finally, we apply our TOPSIS* method on the fuzzy

matrices that have been measured to order our

alternatives. Regarding the alternatives selected for

this study, 5G/6G and DSRC/WAVE wireless

mediums have been selected as the most used means

of communication in a VANET network by the

vehicles moving within it.

Figure 1 : System Model

4 METHODS USED IN OUR

STUDY

4.1 Fuzzy Set Theory

Zadeh [10] introduced the fuzzy set theory to reflect

the uncertainty of human decisions and thoughts. its

ability to represent vague data is a very important

contribution in the field of mathematics, especially

the one related to abstract, vague object classes. A

fuzzy set is represented by a function called

membership function F (1) which associates for any

point X a real number in the interval [0,1]

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

388

(1)

In the pairwise comparison, the TFN defined by the

three real numbers (l,m,h) in order to express a fuzzy

event.

4.2 Method Fuzzy AHP

AHP technique is used to process and analyze

complex decisions, although this technique has

weaknesses in its interpretation and adaptability to

heterogeneous systems. Fuzzy AHP has emerged as a

contouring solution combining AHP with fuzzy logic.

The importance of this method is in the phase of

generating for each pair of factors fuzzy relative

importance. the fuzzy evaluation matrix is thus

obtained:

(2)

Such us:

𝑝



𝑙



,𝑚



,𝑢





Knowing that there are several implementations of

the weighting process via FAHP, we chose the one

proposed by Bucklet [11]. This one uses the

geometric mean approach to calculate the resultant

vector in the pairwise comparison matrix:

(3)

with:

Lastly, we apply the Fuzzy AHP method to each

QoS class, and the associated weight vectors are

generated for each of the criteria.

4.3 Fuzzy TOPSIS

The Technique for Order Preference by Similarity to

Ideal Solution (TOPSIS) was designed in the 1980s.

It is a ranking method that is easy to implement and

apply. It aims at selecting the best alternative that has

the shortest distance to the positive ideal solution and

the farthest distance to the negative solution

simultaneously.

The distances separating each alternative X from the

ideal positive and anti-ideal solutions are given as

follows:

(4)

We calculate just after the relative proximity to the

ideal solution with the following formula:

(5)

4.4 Proposed TOPSIS*

The procedure we followed considers the mobile

feature of the vehicles not taken into account by the

classical FTOPSIS method. For this reason, we

started by modifying the previous equation,

introducing two new parameters (b,w) expressing

the relative importance towards the anti-ideal and

ideal solution calculated by applying FAHP for each

of the QoS classes.

To calculate the new value of the relative

proximity of the optimal solution, we propose the

following equation:

5 SIMULATIONS AND RESULTS

5.1 Simulation

The simulation performed for both 5G and DSRC

wireless technologies involved the four QoS class

types (Streaming, Conversational, Interactive, and

Background) that cover the different user-side

requirements. The Generation of the values for each

of the criteria was randomly produced based on the

ranges specified in Table 1; the simulation was

performed in 1000 vertical handover decision cases

using a java program designed for this purpose.

Using a New Fuzzy MADM Method to Improve the Number of Vertical Handovers in the Case of Wireless Technology Selection for

VANETs

389

Table 1 : QoS Criteria

Tech Latency

(ms)

Throu-

ghput

(

Mb/s

)

Data

rate

(

Mb/s

)

Coverage

(m)

DSRC 100-

1000

0.01-

100-

500

150-

500

5G/

10-

100

0.1-

100

1000-

10^5

10-

100

First, we constructed our pairwise comparison

matrix using the linguistic variables[12] listed in

Table 2, and then we apply the FAHP method to

generate the weights per criteria (Latency,

Throughput, Data Rate and Coverage). Next, we

apply our AHP method again to determine the

importance of b,w relative to the ideal solution and

the anti-ideal solution, respectively (Table 3). Finally,

we apply the enhanced TOPSIS method (TOPSIS*)

to measure the new closeness to the ideal solution and

rank our alternatives for the four classes of QoS.

Table 2 :Membership function of linguistic scale

Fuzzy

Numbe

Linguistic Scales TFN

1̃ Equally important (Eq) (1,1,1)

3̃ Weakly important (Wk) (2,3,4)

5̃ Essentiall

ortant

(

)

(

4,5,6

)

7 Very Strongly important

(Vs)

(6,7,8)

9 Absolutely important (Ab) (9,9,9)

Table 3 : Values of b and w for each QoS class

QoS Class

Streamin

0.720 0.300

Conversational 0.850 0.150

Back

roun

0.750 0.180

Interactive 0.900 0.080

5.2 Results and Discussion

The comparison results presented in this paper show

that the improved TOPSIS method is more efficient

than the conventional TOPSIS method for wireless

technology selection in a heterogeneous and

topologically dynamic environment. In the first

simulation, we evaluate the proposed approach

compared to the classical method, the second

simulation, comparing the average numbers of

Vertical Handover by each methods.

It should be noted that the number of simulations

carried out allowed us to go through the different

situations that a vehicle using a vehicular ad-hoc

network network may face, as well as having a

database of different parameter values (latency,

bandwidth, coverage...) that can give them an

advantage to override or signal an urgent obstruction

on the way.

The following figure shows that the applied

method reduces considerably the number of Vertical

Handover for the four QoS classes (Streaming,

Conversational, Background, Interactive) allowing to

overcome the deficiencies of the classical method.

And to be a suitable solution for autonomous mobile

vehicles evolving in a heterogeneous environment.

Figure 2 :Average of the number of Vertical Handover for

all Qos Class

The new TOPSIS* approach allows, as explained, to

reduce the number of handovers; for example, it

reduces the number of handovers by 7% compared to

the classical Fuzzy TOPSIS method. Below is a table

of the different improvements that the new approach

offers compared to the classical method for the four

QoS classes.

Table 4 :Improvement of the VH of the new approach.

Traffic Class TOPSIS*

Streamin

Conversational 7

Back

roun

Interactive 3

6 CONCLUSION

This paper proposes a new approach to improve the

selection of wireless technologies in the VANET

network by improving the standard Fuzzy TOPSIS

alternative selection method to fit the VANET

context. Our approach uses the Fuzzy AHP method to

TOPSIS*

FTOPSIS

10%

20%

30%

40%

50%

60%

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

390

measure the relative weight of the selection criteria.

The simulation we performed shows that the

proposed method significantly improves vertical

handovers compared to the classical method.

In future work, we intend to improve our java

program used in the simulation to include the routing

protocol settings during inter-vehicle communication

and include other parameters to fit better with the

VANET context.

REFERENCES

S. Ftaimi et T. Mazri, « Handling Priority Data in Smart

Transportation System by using Support Vector

Machine Algorithm », Adv. sci. technol. eng. syst. j.,

vol. 5, n

6, p. 1422‑1427, déc. 2020, doi:

10.25046/aj0506172.

S. Al-Sultan, M. M. Al-Doori, A. H. Al-Bayatti, et H.

Zedan, « A comprehensive survey on vehicular Ad Hoc

network », Journal of Network and Computer

Applications, vol. 37, p. 380‑392, janv. 2014, doi:

10.1016/j.jnca.2013.02.036.

K. Abboud et W. Zhuang, Mobility Modeling for Vehicular

Communication Networks. Cham: Springer

International Publishing, 2015.

E. Stevens-Navarro et V. W. S. Wong, « Comparison

between Vertical Handoff Decision Algorithms for

Heterogeneous Wireless Networks », in 2006 IEEE

63rd Vehicular Technology Conference, Melbourne,

Australia, 2006, vol. 2, p. 947‑951, doi:

10.1109/VETECS.2006.1682964.

M. Lahby, A. Attioui, et A. Sekkaki, « An Optimized

Vertical Handover Approach Based on M-ANP and

TOPSIS in Heterogeneous Wireless Networks », in

Advances in Ubiquitous Networking 2, vol. 397, R. El-

Azouzi, D. S. Menasche, E. Sabir, F. De Pellegrini, et

M. Benjillali, Éd. Singapore: Springer Singapore, 2017,

p. 15‑29.

K. Savitha et C. Chandrasekar, « Trusted Network

Selection using SAW and TOPSIS Algorithms for

Heterogeneous Wireless Networks », IJCA, vol. 26, n

8, p. 22‑29, juill. 2011, doi: 10.5120/3125-4300.

F. Torfi, R. Z. Farahani, et S. Rezapour, « Fuzzy AHP to

determine the relative weights of evaluation criteria and

Fuzzy TOPSIS to rank the alternatives », Applied Soft

Computing, vol. 10, n

2, p. 520‑528, mars 2010, doi:

10.1016/j.asoc.2009.08.021.

V. Sasirekha et M. Ilanzkumaran, « Heterogeneous wireless

network selection using FAHP integrated with TOPSIS

and VIKOR », in 2013 International Conference on

Pattern Recognition, Informatics and Mobile

Engineering, Salem, févr. 2013, p. 399‑407, doi:

10.1109/ICPRIME.2013.6496510.

Y.-M. Wang et Y. Luo, « On rank reversal in decision

analysis », Mathematical and Computer Modelling,

vol. 49, n

5‑6, p. 1221‑1229, mars 2009, doi:

10.1016/j.mcm.2008.06.019.

L. A. Zadeh, « Similarity relations and fuzzy orderings »,

Information Sciences, vol. 3, n

2, p. 177‑200, avr.

1971, doi: 10.1016/S0020-0255(71)80005-1.

J. J. Buckley, « Fuzzy hierarchical analysis », Fuzzy Sets

and Systems, vol. 17, n

3, p. 233‑247, déc. 1985, doi:

10.1016/0165-0114(85)90090-9.

A. Bahnasse, F. E. Louhab, H. Ait Oulahyane, M. Talea, et

A. Bakali, « Novel SDN architecture for smart MPLS

Traffic Engineering-DiffServ Aware management »,

Future Generation Computer Systems, vol. 87, p.

115‑126, oct. 2018, doi: 10.1016/j.future.2018.04.066.

Using a New Fuzzy MADM Method to Improve the Number of Vertical Handovers in the Case of Wireless Technology Selection for

VANETs

391