Computer Networks 183 (2020) 107601

Contents lists available at ScienceDirect

Computer Networks
journal homepage: www.elsevier.com/locate/comnet

Handover scheme for 5G small cell networks with non-orthogonal

multiple access
Ali Çalhan a, *, Murtaza Cicioğlu b
a
Computer Engineering Department, Düzce University, Düzce, Turkey
b
Information Technologies Department, Ministry of National Education, Bolu, Turkey

A R T I C L E I N F O A B S T R A C T

Keywords: The enormous developments in new generation networks reveal the high bandwidth requirements of mobile
Handover users. Therefore, wireless communications technologies move higher frequencies with a small coverage area. The
Fuzzy logic small cell concept exposes these advancements for 5G network framework. The small cell issue is a part of 5G
Small cell
networks with non-orthogonal multiple access, millimeter-wave communications, massive MIMO technology,
5G
NOMA
IoT, spectrum sharing, ultra-dense networks, D2D communications, etc. The small cell with small coverage area
and seamless communications needs of the mobile users reveal the dense networks with explosive data traffic.
The handover process undertakes an important task for 5G networks according to the dense network, small
coverage area, and user mobility specifications. In this study, we aim to investigate classic RSSI based handover
scheme and fuzzy logic based handover scheme with NOMA for 5G networks. We take into consideration RSSI,
BER, and Outage Probability parameters for fast and seamless handover decisions and utilize NOMA technique
for users’ multiple access. This is the first study for 5G small cell networks fuzzy logic-based handover scheme
with NOMA in the literature.

1. Introduction mechanism [3]. Besides, some small cell coverage areas overlap each
other. Sometimes the mobile users may change the serving small cell
Recent developments of new generation wireless networks have base station frequently and unnecessary.
provided a transition from macrocell base stations (2G and 3G) to The signals with higher frequencies cannot pass in walls and blocks,
microcell base stations (5G and 6G). The microcells are called picocells, and also they reflect these obstacles. These problems may cause fluc­
femtocells, and small cells covering 100 s to 10 s of meters in the liter­ tuations in signals from small cell base stations. Received Signal
ature [1]. Especially in 5G networks, the small coverage area as a key Strength Indicator (RSSI) is the basic measurement for signal quality. In
technology gives higher data transfers and provides greater network the nature of the handover process, the RSSI parameter is used for the
capacity for mobile users. The basic idea with the small cell is to get the decision of serving base station. Besides, various parameters can be
access cells as close as possible to the mobile users. These small cells utilized in handover decision phase like Signal to Noise Ratio (SNR) for
access the mobile users with small transmission power, as a result of homogeneous networks as a single parameter, and data rate, monetary
small coverage. Therefore, the small cell base stations may be placed in cost, bandwidth, etc. are used for heterogeneous networks as multiple
the streets, on trees, and lampposts. In the near future, we will come parameters.
across these base stations everywhere. As said before, the aim of the 5G networks is to provide higher data
The main idea under the small cell concept is to increase the reuse of rates in gigabits per second level. This goal cannot be achieved with only
the spectrum for increasing network capacity [2]. It is very easy to small cells and besides, new techniques must be adopted with small cell
implement the small cell network capacity with higher frequencies for such as non-orthogonal medium access, millimeter wave, massive MIMO
immobile users. They do not move during the communications in a small to 5G networks.
cell coverage area. It is a bit complicated for mobile users because they Orthogonal Multiple Access (OMA) techniques are the essential
have to change the small cell frequently during connections. From one specifications for wireless networks until 5G networks. Frequency Di­
small cell base station to another the mobile users move with handover vision Multiple Access (FDMA) and Time Division Multiple Access

* Corresponding author.
E-mail address: [email protected] (A. Çalhan).

https://doi.org/10.1016/j.comnet.2020.107601
Received 21 August 2020; Accepted 11 October 2020
Available online 13 October 2020
1389-1286/© 2020 Elsevier B.V. All rights reserved.
A. Çalhan and M. Cicioğlu Computer Networks 183 (2020) 107601

Fig. 1. The illustrations of the multiple access techniques.

(TDMA) have been used for 2G wireless networks that utilize separate The organization of the paper begins the current section as Intro­
frequency bands and time slots for each user. The desired data rate by duction, then continues the Related Works as Section 2. The NOMA and
mobile users has not too much and the bandwidth can be reserved by handover concepts with the system model are given in Section 3. The
such techniques. With 3G networks, mobile users and their desired data simulations are shown with the results in Section 4. The paper concludes
rates have increased and Code Division Multiple Access (CDMA) tech­ with the last section about recommendations for future work.
nique has been utilized with spreading codes. In CDMA, all users have a
specific code and they can send their data with same frequency bands at 2. Related works
the same time. Then, 4G networks have been used Orthogonal Fre­
quency Multiple Access (OFDMA) by assigning subsets of subcarriers to There are several studies in the literature about NOMA, 5G Net­
mobile users. This allows for simultaneous low data rate transmission works, fuzzy logic based handover management separately. But there is
from several mobile users. These OMA techniques can support limited no study including fuzzy logic-based handover management in 5G small
mobile users because of aforementioned reasons. These techniques are networks that use NOMA technique. Many studies and several problems
illustrated in Fig. 1. with 5G networks have been examined and finally, this study has been
Non Orthogonal Multiple Access (NOMA) is capable of supporting a occurred for constructing a new handover scheme for 5G small cell
massive number of mobile users with higher data rates for 5G networks networks with NOMA.
[4]. NOMA aims to support more mobile users using especially power The NOMA idea was suggested in [5] as a new medium access control
domain. There are three main types of NOMA as power-domain NOMA, (MAC) for new generation networks towards 2020s. Then, the authors
code-domain NOMA, and NOMA multiplexing in multiple domains. discuss the concept and practical considerations of non-orthogonal
Power-domain NOMA supports multiple users within the same time, multiple access with a successive interference canceller (SIC) at the
frequency, code communication resources by distinguishing them with receiver side in [6]. In [7], the authors consider a dual-hop cooperative
different power levels. Code-domain NOMA uses user-specific spreading relaying scheme using NOMA, two sources communicate with their
sequences that are either sparse sequences or non-orthogonal cross-­ corresponding destinations in parallel over the same frequency band via
correlation sequences of low correlation coefficient. The power-domain a common relay. In [8], a cooperative NOMA transmission scheme that
NOMA is considered in our study. uses the fact that some users in NOMA systems have prior information
The NOMA promises better user fairness. Hence it uses power- about the others’ messages has been proposed.
domain multiplexing by using Successive Interference Cancellation There are a few studies can be found in the literature including
(SIC) for multiuser detection. SIC mechanism removes intelligently the NOMA, small cell, 5G, and handover keywords. In [9], the concept of 5G
interference. NOMA uses SIC as a decoding method at the receiver and vehicular small cell networking (5G V-SCN) has been introduced and
uses superposition coding at the transmitter. All users have different focused on interference management and handover issues with NOMA.
power levels in the same physical channel. A hierarchical power control solution to achieve the joint optimization
The frequent user connection transitions among dense networks are of cell association and power control to increase the spectrum and en­
evitable such overlapped small coverage base stations. Handover man­ ergy efficiency in NOMA-enabled 5G V-SCN. In [10], a hybrid MAC
agement aims to improve the effective user connection transfers in consists of NOMA, and OMA has been proposed for 5G Heterogeneous
cellular networks. In this paper, we focus on a small cell handover Networks, and a dynamic power allocation has been used and four
process with power-domain NOMA for 5G networks. The aim of the generic pairing methods for NOMA: Hungarian, Gale-Shapley, random,
study is to decide the smart handover on power-domain users. We and exhaustive have been considered. In [11], Heterogeneous
investigate several simulations according to the RSSI thresholds and ultra-dense networks (H-UDNs) consisting of virtualized integrated
observe the ping-pong effect and bit error rate variations. Also, fuzzy ground-air-space radio access networks (RANs) have been proposed and
logic-based handover management has been developed for 5G small cell network coordination for virtualized H-UDN to efficiently manage
networks. computing resources and inter-cell interference has been studied. The

2
A. Çalhan and M. Cicioğlu Computer Networks 183 (2020) 107601

Fig. 2. The simulation environment.

study includes multiple access (NOMA) to efficiently manage inter-cell user) have connections in the small cell coverage area for downloading
interference and improve network performance. operations. U1 has d1 and U2 has d2 data for communication with small
To the best knowledge of the authors, our study has some differences cells. In the encoding and transmission phases of the power-domain
and novelties as below: NOMA, the data are transferred to the destination with some amount
of power as α1 and α2 with a whole. The data are presented as:
• RSSI, BER, and Outage Probability parameters have been evaluated (√̅̅̅̅̅ √̅̅̅̅̅ )
for handover management with fuzzy logic. D = Pt α1 d1 + α2 d2 (1)
• NOMA MAC with realistic simulations specifications has been per­
In Eq. (1), D is the received data at the near user (U2), and Pt is the
formed with handover process.
transmit power of the data. The processes that are done at the trans­
• RSSI based and RSSI with threshold based handover simulations
mitter called superposition coding for power-domain multiplexing as
have been carried out, and RSSI based handover mechanism has been
shown in Eq. (1). In addition, α1 and α2 are the power allocation factors
compared to our proposed fuzzy logic-based handover scheme.
of the users, respectively. For power allocation factors, some rules are
given below must be provided:
3. System model
α1 + α2 = 1 (2)
The system model consists of two mobile users and two small cells in
the simulation environment as can be seen in Fig. 2. These two mobile α1 > α2 (3)
users are within the coverage area of two small cells. Due to mobility,
In the power allocation factor determination phase, there are two
RSSI values also change instantaneously.
main methods used in a fixed and dynamic manner. In a fixed method,
The mobile users’ connections are transferred one to another
the power allocation factors are not changed during communications. It
frequently, and the network quality of service (QoS) may decrease with
is not realistic because of non-having wireless channel conditions, but it
the ping-pong effect.
can be used for effectiveness of the other methods for comparisons. In
For a small cell base station, as mentioned before there are two
dynamic methods, the channel state information (CSI) are considered for
mobile users as near and far. User 1 (U1) (far user) and User 2 (U2) (near

Fig. 3. NOMA users’ power levels and SIC process.

3
A. Çalhan and M. Cicioğlu Computer Networks 183 (2020) 107601

Fig. 4. First scenario simulation environment. Fig. 5. BER performance of NOMA for various wireless channel models.

more realistic communication. There are a few methods named max- 4. Simulations
sum rate, max-min rate, and proportional fairness for dynamic power
allocations scheduling. The details of the power allocation methods will This section consists of two subsections. In the first subsection, we
be given when they are used in the paper. give the NOMA simulation environment with specifications, and the
The copies of the data from base stations are obtained at the users performance analysis has been made. In the second subsection, we
after passing wireless channels with CSI. In this phase of the NOMA, the examine the handover process with various scenarios and methods
obtained data after propagating through wireless channel can be considering the NOMA environment in the first subsection.
expressed in a general manner as below:
4.1. The proposed NOMA
y1 = H1 ∗ D + N1 (4)
In the first scenario, we test the simulation environment for various
y2 = H2 ∗ D + N2 (5) channel models. There are two mobile users as far (User 1) and near
In Eqs. (4) and (5), y1 and y2 are the received data at each user, N1 (User 2) users in the small cell coverage area as shown in Fig. 4. These
and N2 are the AWGN noise for each user, H1 and H2 are the Rayleigh simulations have been done for verifying the consistency of the simu­
fading coefficients. The received data at the users are copies of the data lation environment. The basic specifications have been used as fixed
with CSI from the base station. Each user extracts the own data from D as power allocation, BPSK modulation, the SNR range from − 5 dB to 5 dB,
can be shown in Eq. (1). SIC is an iterative algorithm that is carried out at and 105 number of Monte Carlo simulations. The distance to the small
the receiver in power-domain NOMA for decoding the superposition cell base stations of the mobile users is 50 m and 100 m respectively.
coding data in the order of decreasing power levels. SIC process repeats In the literature, AWGN, Rician, and Rayleigh channel models have
until all users’ data over. The power relationship and the SIC processes been used widely for modeling wireless communication conditions.
are illustrated as in Fig. 3. AWGN channel adds white Gaussian noise to the data signal and
The SINR1 and SINR2 are the Signal to Inference plus Noise Ratio generally known model for thermal noise in communication channels.
(SINR) values, DR1 and DR2 are the achievable data rates of the U1 and The AWGN channel can be stated by the input-output relationship as
U2 after the SIC process can be written as below: below:
2
|H1 | Pt α1 y=x+n (10)
SINR1 = 2
(6)
|H1 | Pt α2 + σ2 where x is the transmitted data signal, y indicates the output of the
2 channel, and n is a real Gaussian process with zero mean, variance σ2n =
|H1 | Pt α2
SINR2 = (7) E[n2 ], and two sided power spectral density N20 .
σ2
The data signal is degraded by white noise n which has a constant
( )
|H1 |2 Pt α1 spectral density and a Gaussian distribution of amplitude in an AWGN
DR1 = log2 (1 + SINR1 ) = log2 1 + 2
(8) channel. The Gaussian distribution has a probability density function
|H1 | Pt α2 + σ 2
(pdf) given by,
( )
|H1 |2 Pt α2 1
Pawgn (n) = √̅̅̅̅̅̅̅̅̅̅e(− n /2σ ) (11)
2 2

DR2 = log2 (1 + SINR2 ) = log2 1 + (9) 2πσ 2

σ2
In addition, Rician and Rayleigh fading models have to be used for
In the next section, we will explain the details of the NOMA and our
more realistic wireless communications. The Rician fading occurs if
proposed simulation environment with specific parameters. More
there is a line of sight (LOS) between transmitter and receiver, if not,
detailed information about NOMA can be found in [5] and [6].
Rayleigh fading happened.
When the received data signal consists of multiple reflective rays plus
a significant line of sight component, the received envelope amplitude
has a Rician PDF as given in Eq. (12), and the fading is referred to as

4
A. Çalhan and M. Cicioğlu Computer Networks 183 (2020) 107601

Fig. 6. The comparative BER analysis of two power allocation schemes. Fig. 7. The outage probability performance.

Rician fading.
x (−
(12)
x2 +A2/2σ2 )I0(xA/σ 2 )
Prician (x) = e
σ2
I0 is the zero-order modified Bessel function of the first kind, A in­
dicates the peak magnitude of the signal component (A ≥ 0), and x is the
faded signal amplitude (x ≥ 0). In addition, the Rayleigh fading means
that a received multipath signal consists of a large number of reflected
waves with independent and identically distributed (IID) phase and
amplitude. The A component in Rician fading approaches to zero and
the Rayleigh PDF can be expressed as below:
x (− x2/2σ2 )
Prayleigh (x) = e (13)
σ2
In Fig. 5, the comparative performance of NOMA systems for
different wireless channel models are illustrated. In the first scenario the
users have fixed power allocation factors as 0.8 and 0.2 for far and near
user, respectively. As can be seen from Fig. 5, the far user BER perfor­
mance is better than near user due to the high power allocation factor.
And also, in AWGN, Rician, and Rayleigh wireless channel models, the
BER increases, respectively.
In the second simulation model, the NOMA performance has been Fig. 8. Number of handover according to various RSSI thresholds.
tested with Rayleigh fading and a dynamic power allocation method for
more realistic simulation. In the proposed fairness power allocation α1
α1 + α2 = 1
and α2 are dynamically adjusted based on target rate requirement and
α1 > α2 (18)
CSI. In the proposed power allocation scheme, the objective is to α1 > 0
maximize the far user power allocation factor for reaching sufficient
throughput. The total data rate can be written as: Finally, the occurred power allocation factor can be shown:
( )/
DRtotal = DR1 + DR2 (14) α1 = maxDR1 |H1 |2 Pt σ2 |H1 |2 Pt (1 + maxDR1 ) (19)
( ) ( )
2 2
DRtotal = log2 1 +
|H1 | Pt α1
+ log2 1 +
|H1 | Pt α2
(15) α2 = 1 − α1 (20)
2
|H1 | Pt α2 + σ 2 σ2
The BER results are given in Fig. 6 for the second scenario with fixed
In addition, the maximization problem occurs when the desired data PA and fairness PA. As can be seen from Fig. 6, the proposed fairness PA
rate is obtained for far user. For this reason, the power allocation factor gives better BER performance than fixed PA.
of U1 as α1 is found with objective function as below: For a more detailed comparison, the outage probability capabilities
of the fixed and fairness power allocation schemes are given in Fig. 7.
P = Max(α1 ) (16)
Outage probability is defined as the point at which the receiver power
Also, we can give the below formula for objective function: value falls below the threshold, and so, given data rate is not supported
because of variable channel capacity.
maxDR1 = Max(DR1 ) (17)
As can be seen from Fig. 7, the outage probability rises, peaks and
then starts to drop. The power allocation factor optimization of far user
and the constraints can be written as below:
gives better the outage probability about 10 bps/Hz data rate, but the

5
A. Çalhan and M. Cicioğlu Computer Networks 183 (2020) 107601

increased, the number of handover decreases. Fewer handovers are

preferred in dense networks with small cell coverage areas. In addition,
the bit error rate (BER) results are given for User 1 for the first handover
scheme in Fig. 9.

4.2.1. Fuzzy based handover management

In dense wireless environments, only RSSI is not enough for hand­
over management because it causes a high number of handover. Various
parameters must be used in handover decision phase for more realistic
handover management. In 5G small cells with a highly dense network
environment using NOMA MAC, BER and Outage Probability with RSSI
are dominant parameters for selecting a small cell base stations. For
getting BER and outage probability results of 5G small cells with NOMA,
some simulations have been made in the previous sections.
Fuzzy logic is one of the most important and frequently used multi-
attribute decision-making methods in the literature. In Fuzzy, there
are several sets for classification for each parameter, unlike classical set
theory. In set theory, an event is a member of a set or not. In fuzzy logic,
an event can be a member of more than one set with membership value.
A fuzzy system consists of a rule base, fuzzification unit, inference en­
Fig. 9. User 1 BER results. gine, and defuzzification unit. The rule base comprises the set of rules
and the “IF-THEN” conditions provided by the experts to manage the
near user data rate decreases due to the increase of far user power factor. decision-making system, on the basis of linguistic information. The
Then fairness power allocation efforts to balance the power factors, first fuzzification unit converts the system inputs as crisp numbers into the
the outage probability of far user increases, and then begins to decrease. fuzzy sets. The inference engine controls the matching degree of the
current fuzzy input with respect to each rule and selects which rules are
to be fired according to the input field. Then, the fired rules are com­
4.2. The handover management bined to form the control actions. The defuzzification unit converts the
fuzzy sets obtained by the inference engine into a crisp value.
In the second subsection, we have performed various handover The fuzzification processes can map the crisp input values into fuzzy
management schemes with the proposed NOMA environment as given in membership functions. And the opposite defuzzification processes can
Fig. 2. The simulation results have been gathered for User 1 with SNR 10 be used to map a fuzzy output membership function into the crisp output
dB. The first handover scheme is based on the RSSI values of two small value that can be then used for decision or control purposes. In our
cell base stations. The small cell base stations’ RSSI values are changed fuzzy-based handover system, there are three inputs as RSSI, BER,
instantaneously, so handover can happen frequently. For the RSSI based Outage Probability, and one output as Small Cell Candidate Value
handover performance analysis in 5 G Networks with NOMA, some (SSCV). These membership functions are given in Fig. 10.
scenarios have been developed with RSSI values. In simulations with For a numerical example, a small cell has a − 65 dBm RSSI value, 0.17
RSSI value, User 1 makes 460 handovers between small cell 1 (RSSI1) BER, and 0.48 Outage probability, the fuzzy decision system calculates
and small cell 2 (RSSI2) as can be seen from Fig. 8. User 1 prefers the 0.5436 as SSCV. For each small cell base station, the fuzzy logic hand­
small cell with more RSSI value (RSSI1> RSSI2). If there is a threshold in over scheme produces SSCV. The small cell base station that has greater
handover decision, for example, 0.5 dBm (RSSI1>+0.5 RSSI2) the SSCV is selected as a serving base station. We have compared the pro­
number of handover decreases. In Fig. 8, several threshold values have posed fuzzy-based handover mechanism with the aforementioned RSSI
added to the handover decision, and the impact of the threshold can be based handover mechanism. Firstly, we have performed the RSSI based
seen as reducing the number of handovers. As the threshold value is handover, and there are many handovers that happen in a short

Fig. 10. The membership functions of the proposed handover mechanism.

6
A. Çalhan and M. Cicioğlu Computer Networks 183 (2020) 107601

intelligence and multi-attribute decision-making algorithms can be used

for handover processes.

Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

References

[1] D. Muirhead, M.A. Imran, K. Arshad, A survey of the challenges, opportunities and
use of multiple antennas in current and future 5G small cell base stations, IEEE
Access (2016), https://doi.org/10.1109/ACCESS.2016.2569483.
[2] C. Fan, B. Li, C. Zhao, Y.C. Liang, Regret matching learning based spectrum reuse in
small cell networks, IEEE Trans. Veh. Technol. (2020), https://doi.org/10.1109/
TVT.2019.2947265.
[3] T.M. Duong, S. Kwon, Vertical handover analysis for randomly deployed small cells
in heterogeneous networks, IEEE Trans. Wirel. Commun. (2020), https://doi.org/
10.1109/TWC.2019.2963829.
[4] Z. Ding, H. Dai, H.V. Poor, Relay selection for cooperative NOMA, IEEE Wirel.
Commun. Lett. (2016), https://doi.org/10.1109/LWC.2016.2574709.
[5] Y. Saito, Y. Kishiyama, A. Benjebbour, T. Nakamura, A. Li, K. Higuchi, Non-
Fig. 11. The RSSI based handover simulation results. orthogonal multiple access (NOMA) for cellular future radio access, in: IEEE
Vehicular Technology Conference, 2013, https://doi.org/10.1109/
VTCSpring.2013.6692652.
[6] A. Benjebbour, Y. Saito, Y. Kishiyama, A. Li, A. Harada, T. Nakamura, Concept and
practical considerations of non-orthogonal multiple access (NOMA) for future radio
access, in: International Symposium on Intelligent Signal Processing
andbCommunication Systems (ICPACS 2013), 2013, p. 770.
[7] M.F. Kader, S.Y. Shin, M.B. Shahab, Exploiting non-orthogonal multiple access in
cooperative relay sharing, IEEE Commun. Lett. 21 (5) (2017) 1159–1162, https://
doi.org/10.1109/LCOMM.2017.2653777.
[8] Z. Ding, M. Peng, H.V. Poor, Cooperative non-orthogonal multiple access in 5G
systems, IEEE Commun. Lett. 19 (8) (2015) 1462–1465, https://doi.org/10.1109/
LCOMM.2015.2441064.
[9] L.P. Qian, Y. Wu, H. Zhou, X. Shen, Non-orthogonal multiple access vehicular small
cell networks: architecture and solution, IEEE Netw. (2017), https://doi.org/
10.1109/MNET.2017.1600278.
[10] A.S. Marcano, H.L. Christiansen, Impact of NOMA on network capacity
dimensioning for 5G HetNets, IEEE Access (2018), https://doi.org/10.1109/
ACCESS.2018.2799959.
[11] Z. Zhang, G. Yang, Z. Ma, M. Xiao, Z. Ding, P. Fan, Heterogeneous ultradense
networks with NOMA: system architecture, coordination framework, and
performance evaluation, IEEE Veh. Technol. Mag. (2018), https://doi.org/
10.1109/MVT.2018.2812280.

Ali Çalhan received his M.Sc. and Ph.D. degrees from the
University of Kocaeli, Turkey in 2006 and 2011. Since 2011, he
has been a member of the Computer Engineering Department
of Düzce University. His research interests are wireless com­
munications, cognitive radio networks, body area networks,
Fig. 12. The fuzzy based handover simulation results. and software-defined networks.

simulation time as can be seen in Fig. 11.

Under the same conditions, the proposed fuzzy logic based handover
system results are given in Fig. 12. There are a few fluctuations in the
graph because fuzzy gives more effective and general results with more
than one parameter. It is desirable to make as few handovers as possible.

5. Conclusions Murtaza Cicioğlu received his Ph.D. degree in Electrical-

Electronic and Computer Engineering from Düzce University,
In this study, the NOMA MAC for 5G small cell networks is performed Turkey in 2020. Since 2009, he has been a member of the In­
formation Technologies Department of Ministry of National
with the handover mechanism in simulations. The main goal is to Education, Turkey. He is a member of the Software Defined
develop more effective handover management for dense networks with Networks Community, IEEE. His research interests include
minimizing number of handovers. For this purpose, the dominant pa­ software-defined networking, wireless communications, body
area networks, cognitive radio, and Riverbed Modeler
rameters in the small cells as RSSI, BER, and Outage Probability are (OPNET) simulation software.
detected with simulations step by step in the literature. For multi-
attribute decision-making, fuzzy logic is utilized for handover manage­
ment. The results show that various parameters give more effective base
station selection decisions. For future works, multi-users and different
parameters will be used for handover management. May be artificial