Content uploaded by Nitin Gupta
Author content
All content in this area was uploaded by Nitin Gupta on Jan 27, 2020
Content may be subject to copyright.
Fair Sub-Carrier Allocation in OFDMA and
Cognitive Radio Based IEEE 802.22 WRAN
Joshika Agarwal, Rishabh Jain, Sanjay Kumar Dhurandher
CAITFS, Division of Information Technology
Netaji Subhas Institute of Technology
University of Delhi, New Delhi, India
E-mails: {joshika.agarwal, jainrishab1996, dhurandher}@gmail.com
Isaac Woungang
Department of Computer Science
Ryerson University, Toronto, Canada
E-mail: iwoungan@scs.ryerson.ca
Nitin Gupta
Division of Computer Engineering
Netaji Subhas Institute of Technology
University of Delhi, New Delhi, India
E-mail: nitin3041@gmail.com
Abstract—In both orthogonal frequency division multiple ac-
cess (OFDMA) and cognitive radio based IEEE 802.22 Wireless
Regional Area Networks (WRAN), the problem of assigning the
subcarriers to the users in a fairly manner is challenge. This
paper investigates this issue and proposes a solution in the form
of dynamic programming based sub-optimal algorithm. In this
approach, the sub-carriers are initially allocated considering the
equal power distribution. Then, a fair allocation of sub-carriers
is achieved in such a way that comparatively equal data rate
is obtained for all competing users and no user is completely
starved; and the minimum desired sum rate is achieved in case
a low channel gain user has been assigned the sub-carrier. Upon
completion of the sub-channels allocation to all secondary users
(SUs), the power allocation for each user is performed by means
of the water-filling algorithm. Through simulations, our proposed
scheme is shown to outperform a chosen benchmark scheme in
terms of throughput and fairness, chosen as performance metrics.
Index Terms—Cognitive Radio, IEEE 802.22 WRAN, OFDMA,
Fairness, Total data rate.
I. INTRODUCTION
Radio spectrum is very vital for technological innovations
in wireless communications and for the economic growth of a
country [1]. Due to recent advances in wireless technologies
and the increased usage of these technologies in various
applications, the users of the limited spectrum have grown
exponentially in the recent years. In the next few years, it
is expected that global mobile data traffic will grow up to
49 EB/ month, which is nearly a seven times increase over
year 2016 [2]. In contrast, this has also led to an uneven use
or underutilization of a significant portion of the available
spectrum [3]. To remedy to this wastage of resources, the
cognitive radio paradigm was introduced [4], which promotes
the idea that unused licensed spectrum can be exploited
opportunistically using the so-called spectrum holes, thereby
allowing the SUs to have dynamic access to the spectrum
without harming the operation of the primary user of the band.
Thanks to the IEEE 802.22 standard [5], a system is defined
for WRAN which describe how spectrum holes should be
used for various types of applications. For instance, wireless
broadband access in the WRAN to a rural area is provided
through a base station (BS) to a maximum of 255 fixed units of
customer premises equipment (CPE) using TV white spaces,
which remain unoccupied most of the time [6]. These TV
white spaces are used by the BS and CPE (SU) opportunisti-
cally in the absence of TV services (PU). Since the BS serves
a large area and 802.22 operation is unlicensed, the problem
of self-coexistence among the SUs is of great importance. In
the upstream direction (i..e from CPE to BS), the bandwidth
is shared by the CPE units (SUs) through OFDMA. In the
scheme proposed in this paper, we use OFDM as a modulation
and access technique, which provides an efficient solution
to the frequency selective multi-path fading. Indeed, in this
system, the available bandwidth is divided into sub-carriers,
which are further subdivided by OFDMA into groups (so-
called sub-channels), then assigned to different users. Next,
the power allocated to each used sub-carrier is adjusted in
order to guarantee the minimum required quality of service
(QoS).
Despite providing a larger system capacity, such assignment
will only be beneficial to users that have good channel
properties while those users who have poor channel quality
will be neglected. To address this deficiency, a sub-carrier
allocation scheme that provides a fair allocation of sub-carriers
among multiple users is required. This paper proposes a low-
complexity sub-carrier allocation scheme that ensures a fair
allocation of sub-carriers among the SUs in such a way that
all the competing users achieve comparatively equal data rate.
The rest of the paper is organized as follows. In Section II,
some related works about sub-carrier allocation are presented.
In Section III, the system model is discussed in detail. In
Section IV-A, the proposed sub-carrier allocation scheme is
presented, along with the associated power allocation scheme
2018 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS): SCAN: Advances in Software Defined and Context-Aware
Cognitive Networks
978-1-5386-5979-3/18/$31.00 ©2018 IEEE 571
using water-filling method. In Section V, performance of pro-
posed work is evaluated and simulation results are provided.
Finally, Section VI concludes the paper.
II. RE LATE D WOR K
Several sub-carrier (resp. channel) allocation schemes for
cognitive radio networks (CRNs) have been investigated in the
literature. Few representative are as follows. A comprehensive
survey of channel assignment schemes in CRNs (resp. OFDM)
is presented in [7] (resp. [8]). Most of these works rely on the
proven fact [9] that the sum capacity is maximized when each
sub-channel is assigned to the user with the best sub-channel
gain.
In [10], Al-Imari et al. proposed a scheme where the channel
allocation for the uplink data is done on a sub-carrier to sub-
carrier basis so as to maximize the sum rate. In their scheme, a
water-filling algorithm is used and the sub-carrier is allocated
to the user that possesses the higher data rate.
In [11], Nam et al. proposed a joint subcarrier assignment
and power allocation scheme in full-duplex OFDMA networks,
in which an algorithm is devised to iteratively assign the sub-
carriers to each node that has the highest data rate.
In [12], Gao and Cui proposed a two steps sub-carrier
allocation scheme for CRNs. In the first step, the sub-carriers
are allocated to only those users whose data rates are below the
predefined threshold, then the residual sub-carrier allocation is
performed (it should be noted that this step may lead to an
unfair allocation of channels). After the sub-carrier assignment
is completed, the second step consists of performing a power
allocation for each user over the assigned sub-carriers using a
water-filling algorithm.
In [13], Ng and Sung proposed a subcarrier assignment
scheme for CRNs, where a utility function is designed to
assign the sub-carriers to the SUs in a sequential manner.
This assignment prevails only for the (sub-carrier, SU) pair
that has the maximum data rate; and whenever a sub-carrier
is assigned, all the SUs update their power allocation. This
process is repeated until all sub-carriers have been fully
assigned. It should be emphasized that since this method
requires multiple iterations, its complexity also increases when
the number of SUs increases.
In [14], Krishnan et al. proposed a resource allocation
scheme in OFDM-based CR system, in which the power and
sub-channel allocation are optimized through proportionality
rate constraint. In their approach, the power allocation range
is restricted, and the power allocation strategy takes into count
the expected QoS from the SUs, with the goal of reducing the
interference between the PUs and the SUs while minimizing
the power consumption.
In [15], Baghani et al. studied the non-linear distortion
effects of power amplifiers for the uplink sub-carrier and
power allocation, and proposed an approach to minimize the
out-of-band emissions of a non-linear power amplifier which
creates interference to the other users. However, with their
proposed subcarrier assignment scheme, more than one users
may share the same sub-channel.
In [16], Wang et al. proposed a resource management
scheme for OFDM-based CRNs considering the channel un-
certainties between the SUs and the PUs. In general, it is
assumed that the SUs have perfect knowledge of channel gains
from the transmitters of the SUs to the receivers of the PUs.
But, in this approach, the PUs may not always be cooperative
for offering a feedback mechanism for channel estimation. In
this case, BS are assumed to periodically estimate the channel
condition through a feedback path for each user so as to reduce
the channel uncertainty.
In [17], Dai and Wang proposed a clustering-based spectrum
sharing strategy for CRNs which can be used to segregate the
SUs into multiple groups based on their interference degrees
so that. the SUs in different groups can share the same
OFDM sub-channel. An algorithm is devised to maximize
the sum rate of the SUs in each cluster. Although their
approach can improve the sum rate to some extent, the highly
dynamic nature of cognitive network may frequently affect the
formation of clusters. For a very high arrival rate of the PU,
the SUs may have to form the clusters again and again, which
may lead to increased complexity.
In [18], Xu and Li proposed a resource allocation scheme
for CRNs which makes use of a cooperative game where the
SUs relay the data for the PUs in order to gain access to
the spectrum. In their scheme, a Nash bargaining solution is
proposed to allocate the resources among the PUs and the SUs,
using both decentralized and centralized models. But to ensure
the cooperation among SUs, some communication overhead
among the SUs is incurred.
Unlike most of these previous works where only the channel
gain is considered for sub-carrier allocation purpose, this paper
explores also the concept of dynamic programming and central
tendency for the same. The main idea is to allocate the sub-
carrier to that user whose data rate is higher than the achievable
mean data rate of all users. Simulation results shows that the
proposed scheme achieves a fair allocation of sub-carriers,
with every user achieving a comparable data rate. This is in
contrast to most of the above-discussed schemes where the
user having the higher channel gain is allocated more sub-
carriers that others, and the user with the lower channel gain
may not be allocated any sub-carrier. Therefore, this approach
attempts to decrease the number of starved users as much as
possible.
III. SYS TE M MOD EL
The system under consideration is shown in Fig. 1, which
depicts the network structure of a WRAN [19].
In this model, the TV transmitter and wireless microphones
are the PUs whereas the WRAN BS and CPEs are the SUs.
Here, a set of users K={1,2....K}may try to send their
own data or the sensing results to the BS in the case of a
centralized cooperative sensing [20]. The channel bandwidth
(denoted by B) is divided into M={1, . . . , M }sub-carriers.
A user i∈Kcan transmit over a subset of sub-carriers, with
a transmission power of Pi, subject to an individual maximum
2018 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS): SCAN: Advances in Software Defined and Context-Aware
Cognitive Networks
572
Fig. 1. Network Structure of WRAN
power constraint such that PK
i=1 Pi< Pmax, where Pmax is
the maximum power that can be allocated to a channel.
The SINR for the user ion sub-carrier mi.e. m∈Mis
obtained as:
γm
i=Pm
ihm
ii
n0+Pj6=iPm
jhm
ji
(1)
where Pm
iis the transmission power of user ion sub-carrier
m,hji is the channel gain to the user i’s receiver from user j
on sub-carrier m, and n0is the noise power. Initially, the sub-
carrier allocation is performed using equal power distribution.
This is done in order to reduce the complexity of the iterative
water-filling procedure at this step. In addition, performing the
water-filling procedure at this step may increase the burden
to quickly compute the power in case the channel condition
changes.
The rate at which the user ican transmit on the sub-carrier
m∈Mis given by:
Rm
i=Bmlog21 + γm
i
Γ(2)
where Bis the bandwidth of a sub-carrier and Γ =
−ln (5BE R)/1.5is the bit error rate requirement [21].
Now, let A= [amk]M×Kbe the assignment matrix, then
amk =(1,if user k∈Kis assigned the sub-carrier m∈M.
0,otherwise.
(3)
Therefore, the total data rate of the user iis obtained as:
Ri=
M
X
m=1
amiRm
i(4)
Let Rmin be the minimum sum rate to be achieved after
allocating Msub-carriers to Kusers. The problem can then
be formulated as:
M
X
m=1
Rm
i≥Rmin (5)
such that PK
i=1 Pi≤Pmax and ∀k∈K,∃m∈M|amk = 1
and Ri≈Rj, i, j ∈K.
IV. PROPOSED SUB-CARRIER ALLOCATION SCHEME
The proposed sub-carrier allocation scheme relies the idea
of central tendency. Initially, as none of the sub-carriers is
assigned to any user, the first sub-carrier is assigned to the
user achieving the highest channel gain in it. Then, the data
rate of the assigned user is calculated using Equation (4). Till
now, all other users are getting zero data rate as no assignment
has been done yet. Then, the mean data rate (¯x)for Kusers
is obtained as:
¯x=1
K K
X
i=1
Ri!(6)
Next, let’s Dibe defined as:
Di= ¯x−Ri(7)
Then, the next sub-carrier is allocated to the user having
the maximum value of Di. Doing so, we ensure that the
data rate of all users is approximately equal, which leads
to fair allocation. If Diof more than one user is same, the
tie is broken by allocating the sub-carrier to the user that
has the highest channel gain to that sub-carrier. This process
continues and each sub-carrier is assigned on the basis of
the previous allocation of sub-carriers. In doing so, the data
rate of each user iis bring as closer as possible to ¯x. The
drawback of this approach is that the sub-carrier may be
assigned to a user which is not been allocated a sub-carrier for
a while, i.e. which has a maximum value of Di, but the worst
channel gain among all users to that sub-carrier. This allocation
strategy will reduce the overall sum rate and the objective of
achieving PM
m=1 Rm
i≥Rmin will not be met. To overcome
this issue, only the user that has the highest channel gain is
allocated a sub-carrier, providing that the above objective be
met; otherwise, the next user with the highest channel gain is
considered.
A. Proposed Algorithm
The proposed algorithm is given in algorithm 1.
B. Power Allocation Scheme
In the previous section, each user is assigned some disjoint
set of sub-carriers. Now the power allocation is performed
using the water-filling algorithm, which relies on the idea
that more power is allocated when the channel gain is high,
otherwise less power is allocated [22]. Initially, for a user i, the
sub-carrier with highest channel gain is chosen. Then, based
on the highest gain, the power is allocated to each sub-carrier.
Let gibe the channel gain to noise ratio (CNR) based on
the highest channel gain sub-carrier, then the optimal power
allocation for user iis obtained as:
Pi=λi−1
gi+
(8)
where
(x)+=x, if x≥0
0,otherwise
2018 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS): SCAN: Advances in Software Defined and Context-Aware
Cognitive Networks
573
Algorithm 1 SubAssign(K[K],M[M], Rmin )
K[K]: Set of Kusers
M[M]: Set of Msub-carriers
Rmin : Minimum required sum rate
1: Begin
2: INPUT: A finite set of unused sub-carriers M:
{1,2....M}and users K:{1, . . . , K}, Assignment matrix
A
3: Pick user jwith highest h1
j
4: Update a1j= 1,
5: for int i=2toMdo
6: Calculate the data rate of each user using Equation (4)
7: Calculate ¯xusing Equation (6)
8: Calculate Dfor all users using Equation (7)
9: Pick the user “j”with the highest Dand temporarily
allocate the sub-carrier ito user “j”
10: Rmin =Rmin
(M−(i−1))
11: Calculate Pi
j=1 Rj
12: if Rj≥Rmin then
13: Update aj
i= 1,i+ +
14: else
15: go to Step 9
16: end if
17: end for
18: OUTPUT: one to one mapping of Mto K
and λiis the water level, chosen in such a way that PK
i=1 Pi=
Pmax. Next, the power is allocated to the sub-carriers assigned
to next user (by means of the water-filling algorithm as above)
in such a way that power allocated to each user does not
exceed Pi.
V. SIMULATIONS
The proposed scheme is evaluated using the OMNET++
simulator [23]. The considered simulation parameters are
given in Table I
TABLE I
SIMULATION PARAMETERS
Parameter Value
Bandwidth 20MHz
Noise -80dbm
Interference 82.5dbm
Power allocated to
each sub-carrier
30dbm
Target BE R 10(−5)
Simulation time 10000 sec
The channel gain of each user is initialized in a given sub-
carrier randomly. The performance of the proposed scheme is
compared against that of the considered benchmark scheme
[13]. In this benchmark scheme, the sub-carriers is allocated
to the users on the basis of highest channel gain that they get
on a sub-carrier to sub-carrier basis.
First, the number of users is kept constant at 10 and the
number of sub-carriers is varied from 2to 20. The impact
of this variation on the sum data rate in terms of overall
throughput is investigated for both schemes. The results are
captured in Fig. 2. In Fig. 2, it is observed that the proposed
scheme achieves a throughput comparable to that of the bench-
mark scheme. The benchmark scheme slightly outperforms the
proposed scheme due to the fact that in the later, fairness is
given more weight than the throughput that can be achieved.
However, the proposed scheme is still able to achieve the
minimum specified throughput.
0
1.2
2.4
3.6
4.8
6
2 4 6 8 10 12 14 16 18 20
Throughput (MBPS)
Number of Sub-carriers
Throughput
Proposed Benchmark
Fig. 2. Effect of the variation of the number of sub-carriers on the overall
throughput.
Second, the number of users is kept constant at 10 and the
number of sub-carriers is varied from 2to 18. The fairness
achieved by both the schemes is investigated. Here, the Jain’s
fairness index proposed in [24] is adopted, with a fairness
criteria Fgiven by:
F=Pk
i=1 Ri2
KPk
i=1 R2
i
(9)
where Kis the number of users and Riis the total data rate
of user i. The results are captured in Fig. 3 In Fig. 3, it is
observed that the proposed scheme achieves a fairness of 97%
compared to 69% obtained for the benchmark scheme.
Third, the number of subcarriers is kept constant at 21 and
the number of users is varied from from 2to 20. The impact
of this variation on the sum data rate in terms of overall
throughput is investigated for both schemes. The results are
captured in Fig. 4.
In Fig. 4, it is observed that the benchmark scheme outper-
forms the proposed scheme in terms of throughput achieved
because in the proposed scheme, fairness is given more
weight than the throughput achieved. However, with respect to
fairness (Fig. 5), it is observed that the fairness achieved by the
proposed scheme is 97%, which is superior to that generated
2018 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS): SCAN: Advances in Software Defined and Context-Aware
Cognitive Networks
574
0.0000
0.2500
0.5000
0.7500
1.0000
1.2500
2 4 6 8 10 12 14 16 18
Fainess
Number of Sub-carrier
Fairness
Proposed Benchmark
Fig. 3. Effect of the variation of the number of sub-carriers on the fairness.
0.0000
1.2000
2.4000
3.6000
4.8000
6.0000
2 4 6 8 10 12 14 16 18 20
Throughput (MBPS)
Number of Users
Throughput
Proposed Benchmark
Fig. 4. Effect of variation of the number of users on the throughput.
0.0000
0.2500
0.5000
0.7500
1.0000
1.2500
2 4 6 8 10 12 14 16 18 20
Fairness
Number of Users
Proposed Benchmark
Fig. 5. Effect of the variation of the number of users on the fairness
by the benchmark scheme (51%) when the number of users is
varied.
VI. CONCLUSION
In this paper, we have proposed a sub-carrier assignment
algorithm in the WRAN system, with the goal to achieve
fairness in the allocation. Our proposed approach explores the
idea of central tendency to allocate the sub-carriers to the users
in a proportional fashion such that the minimum sum data
rate is achieved. Once the sub-carriers have been allocated
to the users, the power allocation for each user is performed
by using the water-filling algorithm. Simulation results have
shown that in order to attain fairness, some throughput must
be sacrificed. Indeed: (1) even though the fairness is given
more weight in our proposed scheme than the throughput,
it can still achieve a throughput comparable to that of the
benchmark scheme; and (2) the proposed scheme is superior to
the benchmark scheme in terms of achieves the fairness in sub-
carrier/channel allocation. As future work, we plan to assess
the proposed scheme against other benchmark schemes. We
also plan to design a mechanism for analyzing the interactions
and behaviour of the SUs during the channel sharing process,
which might help reducing the co-channel interference, with
the hope to improve the throughput of the proposed scheme,
keeping the fairness in allocation as the main target.
REFERENCES
[1] S. Bhattarai, J.-M. J. Park, B. Gao, K. Bian, and W. Lehr, “An
overview of dynamic spectrum sharing: Ongoing initiatives, challenges,
and a roadmap for future research”, IEEE Transactions on Cognitive
Communications and Networking, vol. 2, no. 2, 2016, pp. 110-128.
[2] Z. Zhou, Y. Jia, F. Chen, K. Tsang, G. Liu, Z Han, “Unlicensed
Spectrum Sharing: From Coexistence to Convergence”, IEEE Wireless
Communications, Vol. 24, No. 5, 2017, pp. 94-101.
[3] B. Kumar, S. Kumar Dhurandher, and I. Woungang, “A survey of
overlay and underlay paradigms in cognitive radio networks”, Interna-
tional Journal of Communication Systems, vol. 31, no. 2, 2018, DOI:
10.1002/dac.3443.
[4] I. F. Akyildiz, W. Y. Lee, M. C.Vuran, S. Mohanty, “A survey on
spectrum management in cognitive radio networks”, IEEE Commun.
Mag., Vol. 46, No. 4, Apr. 2009, pp. 40-48.
[5] IEEE 802.22 Working Group, http://ieee802.org/22/ (Last visited Dec.
20. 2017).
[6] C. R. Stevenson, G. Chouinard, Z. Lei, W. Hu, S. J. Shellhammer, W.
Caldwell, “IEEE 802.22: the first cognitive radio wireless regional area
network standard”, IEEE Commun. Mag., 2009, pp. 130-138.
[7] E. Ahmed, A. Gani, S. Abolfazli, L. J. Yao, S. U. Khan, “Channel
assignment algorithms in cognitive radio networks: Taxonomy, open
issues, and challenges”, IEEE Comm. Surveys & Tutorials, Vol. 18,
No. 1, pp. 795-823, 2014.
[8] F. Shams, G. Bacci, M. Luise, “A survey on resource allocation
techniques in OFDM (A) networks”, Computer Networks, Vol. 65, 2014,
pp. 129-150.
[9] J. Jang and K. B. Lee, “Transmit power adaptation for multiuser OFDM
systems”, IEEE Journal on Selected Areas in Communications, Vol. 21,
No. 2, 2003, pp. 171-178.
[10] M. Al-Imari, P. Xiao, M. A. Imran, R. Tafazolli, “Low complexity
subcarrier and power allocation algorithm for uplink OFDMA systems”,
EURASIP Journal on Wireless Communications and Networking, J
Wireless Com Network (2013) 2013: 98. https://doi.org/10.1186/1687-
1499-2013-98
[11] C. Nam, C. Joo, S. Bahk, “Joint subcarrier assignment and power
allocation in full-duplex OFDMA networks”, IEEE Transactions on
Wireless Communications, Vol. 14, No. 6, 2015, pp. 3108-3119.
2018 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS): SCAN: Advances in Software Defined and Context-Aware
Cognitive Networks
575
[12] L. Gao and S. Cui, “Efficient subcarrier, power, and rate allocation
with fairness consideration for OFDMA uplink”, IEEE Transactions on
Wireless Communications, Vol. 7, No. 5, 2008, p.1507-1511.
[13] C. Y. Ng and C. W. Sung, “Low complexity subcarrier and power
allocation for utility maximization in uplink OFDMA systems”, IEEE
Transactions on Wireless Communications, Vol. 7, No. 5, 2008, p.1667-
1675.
[14] K. V. Krishnan, R. M. Sajith, S. Khara, “Dynamic resource allocation in
OFDM based cognitive radio system considering primary user QoS and
secondary user proportional constraints”, Journal of Communications
Technology & Electronics, Vol. 60, No. 11, 2015, pp. 1269-1275.
[15] M. Baghani, A. Mohammadi, M. Majidi, M. Valkama, “Uplink resource
allocation in multiuser multicarrier cognitive radio networks under
power amplifier nonlinearity”, Trans Emerging Tel Tech. 2017;28:e3162.
https://doi.org/10.1002/ett.3162, 2017.
[16] S. Wang, W. Shi, C. Wang, “Energy-efficient resource management
in OFDM-based cognitive radio networks under channel uncertainty”,
IEEE Transactions on Communications, Vol. 63, No. 9, 2015, pp. 3092-
3102.
[17] J. Dai and S. Wang, “Clustering-Based Spectrum Sharing Strategy
for Cognitive Radio Networks”, IEEE Journal on Selected Areas in
Communications, Vol. 35, No. 1, 2017, pp. 228-237.
[18] H. Xu and B. Li, “Resource allocation with flexible channel cooperation
in cognitive radio networks”, IEEE Transactions on Mobile Computing,
Vol. 12, No. 5, 2013, pp. 957-970.
[19] Y. Sun, H. Hu, Honglin, F. Liu, H. Yi, X. Wang, “Selection of
sensing nodes in cognitive radio system based on correlation of sensing
information”, Proc. of the 4th IEEE Intl. Conference on Wireless
Communications, Networking and Mobile Computing (WiCOM’ 2008),
2008, pp. 1-6.
[20] X. Chen, H-H Chen, W. Meng, “Cooperative communications for
cognitive radio networks: From theory to applications”, IEEE Comm.
Surveys & Tutorials, Vol. 16, No. 3, 2014, pp. 1180-1192.
[21] A. J. Goldsmith S.-G. Chua, “Variable-rate variable-power MQAM for
fading channels”, IEEE Transactions on Communications, Vol. 45, No.
10, 1997, pp. 1218-1230.
[22] T. M. Cover and J. A. Thomas, “Elements of information theory”, 2nd
Edition, 2012, ISBN-13: 978-0471241959, ISBN-10: 0471241954
[23] https://www.omnetpp.org (Last retrieved Dec. 29, 2017).
[24] R. Jain, D-M. Chiu, W. R. Hawe, “A quantitative measure of fairness
and discrimination for resource allocation in shared computer system”,
Technical Report, Eastern Research Laboratory, Digital Equipment Cor-
poration Hudson, MA, USA, 1984.
2018 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS): SCAN: Advances in Software Defined and Context-Aware
Cognitive Networks
576
