Content uploaded by Mohamed Amine Ouamri
Author content
All content in this area was uploaded by Mohamed Amine Ouamri on Aug 14, 2021
Content may be subject to copyright.
978-1-5386-4225-2/18/$31.00 ©2018 IEEE
Resource allocation in a multi-carrier cell using
scheduler algorithms
Zenadji Sylia
Laboratoire de Génie
électrique
Faculté de Technologie
Université de Bejaia
06000Bejaia, Algeria
sylia.zenadji@univ-bejaia.dz
Gueguen Cédric
Laboratoire IRISA
Université de Beaulieu
Rennes1, France
35 042 Rennes
cedric.gueguen@irisa.fr
Ouamri Med Amine
Laboratoire d’informatique
médicale
Faculté de Technologie
Université de Bejaia
06000 Bejaia, Algeria
md-amine.ouamri@univ-bejaia.dz
Khireddine Abdelkrim
Laboratoire de Génie
électrique Faculté de
Technologie
Université de Bejaia, 06000
Bejaia, Algeria
abdelkrim.khired@gmail.com
Abstract— Resource allocations represent a major challenge and
an important step in the design of cellular network. The
constraint of channel conditions can induce unequal spectral
efficiency typically for users, which suffer path losses. However,
for such a constraint, an inequality in terms of throughput
fairness and delay are observed. In order to solve this problem,
we propose in this paper to use the Schedulers algorithm based on
Proportional Fair (PF), which introduces compensation with the
Signal Interference Noise Ratio (SINR). The results obtained
show by the approach proposed in this case achieved us to
provide an optimal solution for the allocation of resources in the
OFDM multi-carrier system. This distribution of resources leads
to a compromise between throughput and fairness in relation to
schedulers and jointly provides a high throughput.
Keywords— Fairness; Orthogonal frequency division multiple
(OFDM); Proportional Fair; Resources; Scheduling; SINR;
Throughput
I. I
NTRODUCTION
The allocation of resources represents a major challenge for
new generation networks such as 4G and 5G; it has a
significant impact on quality of service (QoS) parameters,
which can be cited, throughput, fairness between the UEs and
the probability of delay. It involves assigning subcarriers to
multiple users. On the other hand, the OFDM technique is an
access method which allows the bandwidth to be divided into
several sub-frequency bands. Moreover, each carrier has a low
bandwidth compared to the total band used, which leads to a
high tolerance for multi-path problems. In this context, recent
years of intensive research have given considerable attention to
the allocation of resources for multi-user OFDM networks.
Authors of [1] proposed a resource allocation assuming
complete reuse of the frequency between Macro and Micro
cells. We investigate the outcome of radio resource allocation
for orthogonal frequency-division multiplexing (OFDMA)
based downlink communication. An intelligent distribution of
the subcarriers in an cellular network using scheduling
algorithms was presented by [2]. [3] explores a simple strategy
to distribute equitably the network capacity by applying Round
Robin scheduling. The research in [4] evaluate different
strategies of ressource allocation from different schedulers and
conclude that perfermance comprehensively depend upon
channel imperfection. In [5] the authors concentrates on an
analytical model for the SINR distribution of the scheduled
subcarriers of an OFDMA system with proportional fair
scheduling, but in exchange the system throughput is reduced.
In this work, we propose to study the allocation of resources
using schedulers in a hexagonal macrocell, namely classical
algorithms such as RR (Round Robin) algorithms,
opportunistic algorithms, PF (Proportional Fair) and MaxSNR
(Maximum Signal to Noise). However, due to multipath fading
phenomenon, users can not enjoy the same capacity even if
they occupy the same position and use the same frequency. To
ensure a better throughput, a very important factor is taken into
account named SINR (Signal Interference Noise Ratio). The
approach introduced in this paper explores proposed
schedulers, and evaluates their performance. We focus on the
planning problem by maximizing the capacity of a downlink
OFDM cellular network while ensuring high equity between
the mobiles. The main objective is to compensate the scheduler
algorithms with a weak SINR at the cell boundaries, by
allocating more resources to associate the UEs.
The remainder of this work is organized as follows. System
model is presented in section II. Scheduling algorithm
strategies are given in section III. Section IV presents the
performance evaluation results of the scheduler compared to
the other schedulers. Finally, in section V, we outline the
conclusion of this paper and future work.
978-1-5386-4225-2/18/$31.00 ©2018 IEEE
II. S
YSTEM MODEL
In this work, we consider a hexagonal macrocell system with a
base station at the center. A set of users is uniformly
distributed over the study area as shown in figure 1. On the
other hand, the channel between the base station and the users
suffers from the path loss, multipath fading and shadow fading.
-1000 -800 -600 -400 -200 0200 400 600 800 1000
-800
-600
-400
-200
0
200
400
600
800
x[m]
y[m]
Fig. 1. Simulation scenario
We calculate the propagation path loss dependent on the
distance
d
between base station and mobile
k
using the
following equation [6]:
()
dPL log6,371,128 +=
()
1
In our case, the multi-path fading was modeled using Rayleigh
fading and following the different paths that the transmitted
signal undergoes before arriving at the reception function of
time. The impulse response can be designated of the channel as
follows [7]:
() ()
()
l
t
L
ll
hth
ττδτ
−
−
=
=1
0
,
()
2
Where,
l
τ
and
()
t
l
h
are the corresponding delay and complex
amplitude of
l
path respectively.
Apply the Fourier transform of (2); we obtain the frequency
response of the time-varying channel at time
t
:
() ()
l
i
et
L
ll
hftH
ω
τ
−
−
=
=1
0
,
()
3
Shadows fading are generated with log-normal distribution and
the random variations in signal amplitude follow a Gaussian
distribution with mean of zero and standard deviation of σ[6].
III. S
CHEDULING
A
LGORITHM
In the cellular networks access to the channel is determined by
the scheduler, they exploit the information on the channel state
CSI (Channel State Information); this information is derived as
a function of the channel gain, the interference conditions and
the estimation of the SINR which is sent by the users to the
base station [8].
The Scheduling is currently hosted in the MAC layer to
optimize resource allocation. The resources are presented in a
time-frequency grid divided in a number of RB resource
blocks [9]. In the frequency domain, the OFDM system
bandwidth is 180 kHz comprising of 12 subcarriers in each
Block Resource (RB) of 15 kHz, in the time domain the
resources are spaced by 0.5ms slots, as shown in Figure 2. In
LTE, the bandwidth value is 1.4 MHz up to 20 MHz, with the
number of RBs ranging from 6 to 100 RBs depends on the size
bandwidth.
Fig. 2. Illustration of a Scheduling block.
Our investigation relies on the use of three different algorithms
namely Round Robin, Proportional Fair and MaxSNR. The
Round Robin scheduler allocates resources in an equally fair
TTI regardless of channel conditions but offers low
throughput. The second algorithm introduced is MaxSNR
algorithm. His principle is to maximize the number of bits
which can be transmitted during a time interval on the
subcarrier n to the mobile
k
, designated by
nk
R,
. In addition,
MaxSNR gives priority to mobiles with the highest signal-to-
noise ratio [10].
()
KkRk
nk
,...,1;maxarg
,
==
∗
()
4
Where,
K
is the total number of active users.
978-1-5386-4225-2/18/$31.00 ©2018 IEEE
Proportional fair is an opportunistic algorithm derived from
MaxSNR. Indeed, systems using PF adapt to the variation in
channel conditions that allows mobile phones to have the same
probability of access to resources [10]. Principle of the
Proportional Fair is sharing in slots and during each slot
()
t
on
the subcarrier n, the base station serves given group of
user
∗
k
, the base station relies on a kind of feedback from the
mobiles to obtain estimate of the channel quality which is in
terms of requested data rate
()
tR
nk ,
for each user
k
, then it
monitors the average throughput
()
tT
nk ,
of
k
th user on n
subcarriers in a window of length
tc
[11].
The objective function representing the PF algorithm is:
()
()
Kk
t
nk
T
t
nk
R
k,....,1;
,
,
maxarg ==
∗
()
5
()
tT
nk ,
is updated for scheduling as in (6),
()
() ()
()
∗≠+
∗=++
=+
kkt
nk
T
c
t
kkt
nk
R
c
t
t
nk
T
c
t
t
nk
T
;
,
1
1
;
,
1
,
1
1
1
,
()
6
The instant service rate on the nth sub-carrier at tth of each
Transmit Time Interval (TTI) is got by [12]:
() ( )
SINRwt
nk
R+= 1
2
log
,
()
7
Where
W
is the total bandwidth.
The received SINR for user
k
on sub-carrier n can be
expressed by:
=+
=M
mN
P
nk
G
km
P
nk
G
nk
P
nk
SINR
1,,
,,
,
()
8
Where,
()
t
nk
P,
,
()
tG
nk ,
are the power assigned by the
serving cell and channel gain on user
k
at subcarrier n
respectively, expressed as a function of the Path loss,
Multipath Fading and the Shadow Fading,
km
P,
is the power
for user
k
on the same subcarrier m,
N
P
is the noise power
spectrum density of AWGN
Predictive PF Scheduling
1: generate different channel responses
2: initialize average throughput for all users
nk
T
,
.
3: for slot t=1 to T do
for carrier n=1 to N do
Compute
()
=+
+= M
mN
P
nk
G
km
P
nk
G
nk
P
wt
nk
R
1,,
,,
1
2
log
,
()
()
t
nk
T
t
nk
R
k
,
,
=
∗
if
()
()
()
∗
=k
t
nk
T
t
nk
R
max
,
,
then
() () ()
t
nk
R
c
t
t
nk
T
c
t
t
nk
T,
1
,
1
1
,++=
else
() ()
t
nk
T
c
t
t
nk
T,
1
1
,
+=
end
end
end
IV.
S
IMULATION
RESULTS
System level simulations were performed to evaluate the
performance of the modified scheduler in terms of downlink
Throughput Fairness and Delay. System level parameters are
summarized in Table 1. In order to compare of the
Proportional Fair algorithm and the basic algorithms, we
designed 800m×800m as area for simulation. However, the
total number of Users is 30; all users are randomly distributed
according to a uniform law and 300 subcarriers. Furthermore,
the approach studied must satisfy a high quality standard,
namely a very high throughput, fairness between users and
delay.
The simulation environment is calculated the matrix of two
dimension with Number users equipment and Carrier Resource
to find the served user every subcarrier, we only simulate the
different scheduling algorithms for different users. The
obtained results are compared between three algorithms, such
as Round Robin, Proportional Fair and MaxSNR for downlink
transmission. We note that the simulation is realized using
MATLAB 2014 and run on PC with Intel core i5 processor @
2GHZ with 4GB RAM.
978-1-5386-4225-2/18/$31.00 ©2018 IEEE
TABLE
I.
P
ARAMETE RS
E
XPERIMENTATION
Parameters
Values
Channel Bandwidth
5MHz
Carrier Frequency
.4Ghz
Number of subcarrier
300
PN
-174 dBm/Hz
Grid layout
1 cell hexagonal
Cell radius
800m
BS Transmitter Power
43 dBm
Channel Model
ITU Pedestrian B
Slot duration
0.5ms
Number of sub-frames
100
Number of users
5,10,15,20,25,30
tc
20
The performance of algorithms is tested. Figure 3 show the
distribution of 300 subcarriers for users operating with a
bandwidth of 5MHz and 100 time slots. The simulation results
show that the Round Robin scheduler allocates these resources
to all users in each subcarrier group which does not account for
multiuser. MaxSNR allocates a different number of subcarriers
for each user to maximize the average throughput. The PF
Scheduler tries to strike a balance between fairness and
maximization of throughput, with almost equitable allocation
of resources between users.
50 100 150 200 250 300
0
5
10
15
20
25
30
35
Number of s ub-carriers
Users
RR MaxSNR PFS
Fig.3. Allocated RB for each user using MaxSNR, PFS and RR.
We have compared the performance of the MaxSNR, PF and
RR for simulated measurements. We concluded according to
Figure 4 that the MaxSNR outperforms the two schedulers in
terms of throughput compared to the other two schedulers,
because this algorithm allocates resources to users with a
strongest channel, a higher number of users, and it maximizes
the system throughput. PF reaches a good level of flow of the
system without Fairness because it exploits the propagation
channel condition. RR reaches the lowest value and remains
constant because this algorithm allocates the resources
independently of users channel response and rate requirements.
510 15 20 25 30
0.5
1
1.5
2
2.5
3
3.5 x 10
7
Users
Throughput (bps)
System Capacity
MaxSNR
PFS
RR
Fig. 4. System
throughput for each user using MaxSNR, PFS and RR.
978-1-5386-4225-2/18/$31.00 ©2018 IEEE
We also studied the influence of multi-user diversity on the
Fairness of the system. The results of simulations plotted in
Figure 5 show that RR and Proportional Fair schedulers
ensures total fairness between users, but MaxSNR is unfairness
algorithm. The delay in the time intervals, according to these
algorithms, is equal to the number of users in the system. The
greater the number of users in the system, the delay is
important, so if we want to achieve quality of service, our
system will only accept users with a delay constraint, PF
reaches zero probability faster, and it has the best behavior
over other algorithms (see Figure 6) because the allocated
subcarriers are sufficient to satisfy the needs of all users.
510 15 20 25 30
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
Users
Fairness
Sys tem Fairness
MaxSNR
PFS
RR
Fig.5. System fairness for each user using MaxSNR, PFS and RR.
510 15 20 25 30
0
0.2
0.4
0.6
0.8
1
delay
probability
Probability delay
MaxSNR
PFS
RR
Fig. 6. Probability delay.
V.
CONCLUSION
In this paper, three schedulers, namely MaxSNR, PF and RR,
have been developed to exploit the idea of multi-user diversity
by taking into account channel conditions in a multi-carrier
cellular network. The three algorithms are tested with both
simulated and real measurements coming from one base
station. The results obtained show the superiority of the
algorithm Proportional Fair in terms of improvement, it makes
a compromise between Fairness and the throughput of the
system.
The future work could include a new service quality
measurement tool which is the Packet Interrupt Ratio (PDOR)
for a more efficient resource allocation.
REFERENCES
[1] Qian Li, Rose Qingyang Hu, Yi Qian,Geng Wu, “Intracell Cooperation
and Resource Allocation in a Heterogeneous Network Relays,” IEEE
transactions on vehicular technology vol. 62, No. 4, May 2013 pp.1770-
1784
[2] Bin Liu, Hui Tian, Lingling Xu, “An Efficient Downlink Packet
Scheduling Algorithm for Real Time Traffics in LTE Systems,” IEEE
CCNC Wireless Communication Track, Vol.10.2013. pp. 364-369
[3] V. Hassel, M.R.Hassen, G.E.oien, “Spectral efficiency and fairness for
opportunistic round robin scheduling, ” IEEE International Conference
Communication in Processing. 2006. Pp. 784-789
[4] Ankit Saxenal, R. Sindal, “Strategy for Ressource Allocation in LTE-
A,” International Conference on Signal Processing, communication,
Power and Embedded system. 2016
[5] D. Parruca, M. Grysla, S. Gortzenand, J. Gross, “ An alytical Model of
Proportional Fair Scheduling in Interference-limited OFDMA/LTE
Networks,” IEEE Vehicular Technology Conference. 2013
[6] 3GPP TR 36.814 V9.0.0, “Further advancements for E-UTRA physical
layer aspects,” Tech. Rep., Mar. 2010
[7] Ye (Geoffrey) Li, Leonard J. Cimini, and Nelson R.
Sollenberger,“Robust channel estimation for OFDM systems with
rapid dispersivefading channels,” IEEE Trans. on Communications,
vol.46, No.7, July1998
[8] S. M. Chadchan, “A fair downlink scheduling algorithm for 3GPP LTE
networks”. International Journal of Computer Network and Information
Security, 6, pp34–41.2013
[9] 3GPP TR 36.815 V9.0.0, “Further advancements for E-UTRA; LTE-
Advanced feasibility studies in RAN WG4,” Tech. Rep., Mar. 2010
[10] D. Parruca, J. Gross, “Throughput Analysis of Proportional Fair
Scheduling for Sparse and Ultra-Dense Interference-Limited
OFDMA/LTE Networks,” IEEE Transactions on Wereless
Communications, vol.10, No.15, October 2016. pp. 6857-6870
[11] P. Viswanath, D. N. C. Tse, and R. Laroia, “Opportunistic beamforming
using dumb antennas,” IEEE Trans. Inf. Theory, vol. 48, no. 5, pp.
1277–1294, Jun. 2002
[12] Zukang Shen, Jeffrey G. Andrews, and Brian L. Evans,
“Adaptive resource allocation in multiuser OFDM systems with
proportional rate constraints,”IEEE Trans. on Wireless
Communications, vol.4, No .6, November 2005
