ArticlePDF Available

Abstract

Traditional cellular networks are moving towards heterogeneous cellular networks (HetNets) to satisfy the stringent demand for data rates and capacity. To enable the new applications in 5G, such as haptic communications, we face new challenges of achieving high availability with low latency in HetNets. In this paper, we introduce coordinated multi-point (CoMP) and carrier aggregation (CA) techniques in HetNets to guarantee the availability of all UEs, where CoMP improves the single-path availability, and CA enhances availability via multi carrier gain combining. To characterize the availability, we first derive an exact closed-form expression for the availability of a random UE in a CoMP and CA-enabled HetNets. To achieve the maximum UE availability, we formulate a maxmin optimization problem. To solve it, we then propose a twostep optimization algorithm and a joint two-step optimization algorithm. The two-step optimization algorithm (TSOA) is based on heuristic algorithm for the optimal subcarrier assignment and UE association, and based on Lagrangian dual method for the power allocation. The joint two-step optimization algorithm (JTOA) is based on genetic algorithm (GA) to achieve the interaction between the first step and the second step. Our results showcase the effective of our proposed JTOA, and the effective of CoMP in availability improvement in HetNets.
1
Availability Analysis and Optimization in
CoMP and CA-enabled HetNets
Jie Jia, Member, IEEE, Yansha Deng, Member, IEEE, Jian Chen, Member, IEEE,
A. Hamid Aghvami, Fellow, IEEE, and Arumugam Nallanathan, Fellow, IEEE
Abstract—Traditional cellular networks are moving towards
heterogeneous cellular networks (HetNets) to satisfy the strin-
gent demand for data rates and capacity. To enable the new
applications in 5G, such as haptic communications, we face
new challenges of achieving high availability with low latency
in HetNets. In this paper, we introduce coordinated multi-point
(CoMP) and carrier aggregation (CA) techniques in HetNets to
guarantee the availability of all UEs, where CoMP improves
the single-path availability, and CA enhances availability via
multi carrier gain combining. To characterize the availability, we
first derive an exact closed-form expression for the availability
of a random UE in a CoMP and CA-enabled HetNets. To
achieve the maximum UE availability, we formulate a max-
min optimization problem. To solve it, we then propose a two-
step optimization algorithm and a joint two-step optimization
algorithm. The two-step optimization algorithm (TSOA) is based
on heuristic algorithm for the optimal subcarrier assignment
and UE association, and based on Lagrangian dual method for
the power allocation. The joint two-step optimization algorithm
(JTOA) is based on genetic algorithm (GA) to achieve the
interaction between the first step and the second step. Our results
showcase the effective of our proposed JTOA, and the effective
of CoMP in availability improvement in HetNets.
Index Terms—Heterogeneous cellular networks, high availabil-
ity, coordinated multi-point, carrier aggregation, genetic algorith-
m.
I. INTRODUCTION
In the past, cellular networks have mainly focused on
achieving higher data rates and greater user capacities for
human-centric applications, such as telephone, mobile inter-
net or video streaming. However, it is expected that future
wireless network will be complemented by a wide range
of innovative and unconventional services and applications,
such as M2M communication, IoT applications, machine-
type communications (MTCs), and haptic communications.
According to Gartner [2], about 6.4 billion connected things
Copyright (c) 2017 IEEE. Personal use of this material is permitted.
However, permission to use this material for any other purposes must be
obtained from the IEEE by sending a request to pubs-permissions@ieee.org.
Manuscript received Aug. 31, 2016; revised Dec. 30, 2016; accepted Feb.
24, 2017. The associate editor coordinating the review of this manuscript and
approving it for publication was Prof. Tony Q.S. Quek. This paper has been
published in part at IEEE International Conference on Communications (ICC),
Paris, 2017 [1].
Corresponding author: Yansha Deng (Email: yansha.deng@kcl.ac.uk.)
Jie Jia and Jian Chen are with the Key Laboratory of Medical Image
Computing of Northeastern University, Ministry of Education, Shenyang
110819, China, and with the School of Computer Science and Engineering,
Northeastern University, Shenyang 110819, China, and also with Department
of Informatics, King’s College London, London WC2R 2LS, UK (email:
{jiajie, chenjian}@mail.neu.edu.cn).
Y. Deng, A.H. Aghvami, A. Nallanathan are with Department of Informat-
ics, King’s College London, London WC2R 2LS, UK (email: {yansha.deng,
hamid.aghvami, arumugam.nallanathan}@kcl.ac.uk).
will be in use worldwide this year, which is nearly up to 30
percent increase from last year. And they also foretasted that
the number would keep growing, and reach nearly 21 billion
by the year 2020.
Usually, these specialized, mission-critical applications have
specific requirements, which seems too stringent for the
conventional human-centric applications, like the ultra high
availability and ultra low latency requirements. Examples of
application having these requirements include haptic com-
munication [3], cloud computing [4], smart energy grids
[5], vehicular communication [6], or factory automation [7].
The temporal availability requirement of these applications
is 99.9999% (six nines) or higher. A more detailed example
is that the factory automation application in a smart factory
needs the end-to-end latency with 1ms and the availability
requirement as 9 nines [8]. In other words, only one message
in 109data transfers can be lost or delayed in more than
1ms. A detailed analysis on future application as well as high
availability requirement can be found in [9]. How to provide
an availability of six nines or even higher for applications with
ultra-reliable requirement using the existing or future cellular
networks has become a major challenge.
The traditional cellular networks are undergoing a signifi-
cant transition to handle the increasing wireless data demands,
as well as the ever increasing availability requirement. Simply
deploying more macro base stations (BSs) is no longer a
sustainable solution to cope with those stringent requirements.
Therefore, action is being taken to deploy more inexpensive,
low-power, small-scale BSs, such as pico, femto BSs, un-
derlaying the conventional cellular networks to improve the
spectral efficiency and reduce the communication distance
[10]. This is the so called heterogeneous cellular networks
(HetNets). However, due to the heterogeneous deployments of
those low power nodes, the interference management among
tiers becomes very challenging and extremely important.
According to the reliability theory [11], there are two
feasible methods to improve availability of a system: 1) the
serial approach, which substitutes or adds more reliable sub-
components in serial with the single sub-component system;
and 2) the parallel approach, which enables multiple sub-
components working in parallel. Multi-connection is one ex-
ample using parallel approach, where the receiver is allowed
to be served by multiple transmitters simultaneously, in or-
der to achieve the multi-link diversity gain and availability
improvement [12]–[16].
In [12], it was shown that the higher availability could be
achieved via multiple less reliable links connection than single
powerful link connection. However, this work was limited to
2
Rayleigh-fading links. The model in [12] has been extended
to [13] by including selection combining and maximal ratio
combining over Nakagami-m fading. In [14], combined macro-
and uplink connections was studied under Nakagami fading
and log-normal shadowing. In [15], multi-input multi-output
(MIMO) was proposed to achieve ultra-reliable and low-
latency communications in 5G machine-type communication
(MTC) use cases. It is noticed that [12]–[15] only considered
the single-user case, and hence the effects resulted from multi-
user and multi-cell interference are not explicitly studied [16].
Due to the different achievable capacity of each link and
cumulative interference caused by all the simultaneously trans-
mitting nodes, nearby or faraway [17], simply considering the
received power from the desired transmitter may not accurately
capture the availability characteristics. A more appropriate
model is to measure the signal quality in terms of the signal-
to-interference-plus-noise ratio (SINR) value. The SINR is,
however, affected by many variable factors that are intrinsic for
any wireless systems. The desired signal power can be slightly
attenuated as a result of the fading nature of the wireless
channel, whereas the received interference can be large due to
the typical aggressive reuse of the time-frequency resources
for maximizing the system capacity in the network. Assuming
the shadowing fading as a random variable, [18] studied the
high availability in wireless networks with different transmit
power at the BS based on SINR model. However, modeling
and analyzing the availability in HetNets based on SINR model
can be computationally and analytically challenging.
Another example of multi-connectivity is the carrier ag-
gregation (CA) technique [19], where the concurrent utiliza-
tion of multiple component carriers (CCs) at the physical
layer is enabled, bringing wider effective bandwidth. CA
has already been included in the 3rd Generation Partnership
Project (3GPP), which combines contiguous or non-contiguous
spectrum fragments to create a virtual wideband channel. The
aggregated bandwidth as large as 100MHz can be obtained by
aggregating 5 20MHz CCs, and the propagation characteristics
of different component carriers may also vary significantly.
e.g., a CC in the 800MHz has very different propagation
characteristic from a CC in the 2.4 GHz. With CA, another
dimension of diversity can be achieved via carrier-selecting
fading [20].
Even though CA improves the availability of single UE in
HetNets via multi-path connectivity, the single path availability
is still low due to the full spectrum reuse in HetNets [21],
[22]. The intercell interference is the most detrimental factor
impairing the single path availability. To cope with it, coordi-
nated multi-point (CoMP) was proposed in 3GPP [23]. With
CoMP, the BSs in coordinated BS cluster are connected via
a backhaul link, and transmit data to the UE simultaneously
to improve single path availability under universal frequency
reuse [24].
Different from CA that exploiting the spectral diversity gain,
CoMP is a coordination technique between BSs to enhance
inter-cell interference coordination in the same carrier. There
are two categories of CoMP: 1) CoMP coordinated beamform-
ing (CoMP-CB) was proposed to avoid inter-cell interference
[25]; and 2) CoMP joint processing (CoMP-JP) is capable
of converting the inter-cell interference to useful signal [26],
therefore provides an external capacity gain over CoMP-CB.
More importantly, with the help of CoMP-JP, the availability
gain of each path can be achieved to bring substantial overall
availability improvement of single UE [27]. However, their
potential to provide ultra-reliable communications in HetNets
has not been treated until recently.
The resource allocation is an effective way to optimize the
system performance for cellular networks with CA/CoMP. In
[20], the joint downlink and uplink resource allocation for
energy-efficient CA was studied, where the optimal power
allocation was solved using the Lagrangian dual method.
In [28], a joint clustering and resource allocation problem
for ultra dense cellular networks was solved using a two-
step algorithm, which consists of the user-centric clustering
based on load information and scheduling algorithm based on
graphic coloring. In [29], the subcarrier allocation algorithm
was proposed to satisfy both throughput and fairness. In
[30], with the aim of maximizing the energy efficiency, the
joint optimization on BS coordination, user scheduling, data
rate adaption, and power allocation was solved via iterative
algorithm.
Different from aforementioned works, the aim of this work
is to achieve high UE availability in HetNets with the help of
CA and CoMP. To the best of our knowledge, this is the first
work taking into account the SINR-based availability modeling
and optimization in HetNets. The main contributions of this
paper are summarized as follows:
We propose CoMP and CA-enabled HetNets to improve
the UE availability in HetNets, where CoMP improves
the availability of each path via interference coordination,
and CA improves the UE availability via multi-carrier
gain combining. This approach for availability improve-
ment is different from previous works only using parallel
method [12], [13].
We present an analytical model for the UE availability in
CoMP and CA-enabled HetNets based on SINR model.
Different from the coverage model or outage model
defined in [31] and [32], where a UE connects to one BS
offering the highest instantaneous SINR, we assume each
UE can be served by multiple subcarriers and multiple
BSs via muli-path connectivity.
We derive an exact closed-form expression for the avail-
ability of a random UE in CoMP and CA-enabled
HetNets, which is verified by Monte Carlo simulation.
Its numerical results reveal the importance of the UE
association, the subcarrier assignment and the power
allocation in achieving high availability.
We formulate an optimization problem with the aims of
maximizing the minimum UE availability for multi users
under the BS transmit power constraint in a multi-cell
HetNet. Due to the complex topology of proposed model,
the optimization problem is NP-hard in nature.
We first propose the two-step optimization algorithm (T-
SOA) to solve the optimization problem. In this algorith-
m, the optimal subcarrier assignment and UE association
is determined first via heuristic algorithm under equal
power allocation, and the optimal power allocation is
3
obtained via Lagrangian dual based method with the
predetermined optimal subcarrier assignment.
We then propose the joint two-step optimization algorith-
m (JTOA), which is an integration of the TSOA with the
GA algorithm to achieve the interaction between the first
step and the second step. Numerical results show that our
JTOA is an effective way for availability improvement.
The remainder of this paper is organized as follows. In
Section II, we present the multi-tier multi-band availability
model. Next, in Section III, we formulate the availability
optimization problem. In section IV, we propose the TSOA and
JTOA. Section V presents the numerical results and Section
VI highlights our conclusions.
II. SYSTEM MODEL
A. Availability Analysis
We consider HetNets including macrocells, picocells, femto-
cells, and further radiating elements. In this network, there are
Nrandomly distributed UEs, denoted as N={1,2, ..., N},
and the set of BSs as B={1,2, ..., S}, where Srepresents
the number of BSs. Each BS has Qavailable bands (e.g.,
800MHz, 2.4GHz, ...), each spectrum band contains For-
thogonal subcarriers. We denote the set of bands in each BS
as Q={1,2, ..., Q}, and the set of subcarriers at each BS
as M={1, . . . , F
| {z }
band1
, ..., (Q1)F+ 1, . . . , QF
| {z }
bandQ
}. We focus on
the downlink transmission with open access for all the small
cells.
With CoMP, the coordinated BS cluster can transmit data
to the UE simultaneously to eliminate the co-channel inter-
ference and improve the received signal quality. The BSs in
each coordinated BS cluster are connected via high-capacity
backhaul links on which complex signaling and user data
are exchanged [33]. To fully exploit the macro diversity of
coordinated BS cluster, we adopt the user-centric adaptive
clustering [34], where any UEs can be served by arbitrary
number of coordinated BSs, and the coordinated BS cluster
for each UE can overlap with each other. Each UE can access
to different sub-carriers in each BS simultaneously, and can
potentially aggregate data from all the available sub-carriers.
For clarify, Fig. 1 presents the illustration of our system model
with 3 BSs and 2 UEs, where each UE can aggregate 2 sub-
carriers, and each UE can also apply CoMP technique on one
sub-carrier to enhance its availability.
To specify the UE association and the resource assignment,
we denote vm
s,n as the resource-allocation indicator binary
variable. If vm
s,n = 1, it indicates that the mth subcarrier of
the sth BS (s∈ B)is allocated to the nth UE (n N ),
and if otherwise vm
s,n = 0 (m∈ M). To specify the power
allocation, we denote the power allocated at the mth subcarrier
of the sth BS is Ps,m, where
Ps,m ((0, P max
s],If UE occupied mth subcarrier of sth BS,
= 0,If no UE occupied mth subcarrier of sth BS,
(1)
and Pmax
sis the maximum transmit power of the sth BS.
s1
Carrier
Aggregation
Carrier
Aggregation
Interference
CoMP CoMP
s2s3
UE1UE2
Marco BS
Pico BS Pico BS
Fig. 1: System model with CoMP and CA
Due to the hardware limitation, the following per-subcarrier
assignment constraint and per BS power constraint need to be
satisfied:
1) The variable vm
s,n must satisfy that each subcarrier for a
BS can only be occupied by at most one UE.
2) The total transmit power at each BS over all its sub-
carriers PmPs,m should not exceed its maximum power
Pmax
s.
We define the set of coordinated BS cluster serving the
nth UE at the mth sub-carrier as Bm
n={s|vm
s,n = 1, s
B, m ∈ M}, the SINR of the nth UE at the mth subcarrier
with CoMP-JP is formulated as
SI N Rm
n=P
i∈Bm
n
Pi,mHi,n Cqdαq
i,n
X
j∈B−Bm
n
Pj,mHj,n Cqdαq
j,n
| {z }
Im
n
+N0
,
(2)
where q=dm/F e, with d·e as the ceiling function, Im
nis
the aggregate interference at the nth UE from all the other
non-coordinated BSs, αqis the path loss exponent of the qth
band, Hi,n is the random variable capturing the fading effects
of the subcarrier between the ith BS and the nth UE, di,n is
the distance between the ith BS and the nth UE, N0is the
noise power, and Cqis the constant depends strongly on carrier
frequency with Cq= (µq
4π)2for the wavelength µq. Similar as
[22], we ignore shadowing and only consider Rayleigh fading
with Hi,n exp(1) for simplicity.
According to the user-centric adaptive clustering, the choic-
es and the size of coordinated BS cluster for each UE can be
adaptive. It can be seen from (2) that increasing the number of
BSs in the coordinated BS cluster for a given UE increases its
SINR value, due to that the coordinated BS cluster convert the
intercell interference between each other to the useful signal.
However, it should be also be noted that increases the number
of BSs involved in this coordinated BS cluster decreases the
probability of other UEs accessibility, due to the fact that all
sub-carriers in a cell are orthogonal and each sub-carrier can
only be occupied by at most one UE at a time.
The signal of a single path cannot be successfully received
if the SINR value SINRm
nis below a certain threshold τ, thus
4
the availability of the nth UE associated with the coordinated
BS cluster Bm
nat the mth subcarrier is expressed as
Am
n=P(SI N Rm
n> τ).(3)
And the availability nth UE in HetNets is defined by
the combination of multiple single-path availabilities (parallel
model), which is derived in the following theorem.
Theorem 1. The availability of the nth UE in a HetNet with
CoMP-JP and CA is derived as
An= 1 Y
m∈M
(1 Am
n),n N ,(4)
where
Am
n=
0 if Bm
n=
Bm
nP
s∈Bm
n
eΞsτN0
ΘBm
n/s
if Bm
n6=, Im
n= 0
BP
s∈Bm
nP
j∈B−Bm
n
eΞsN0τ
jsτB−Bm
n/j ΘBm
n/s
if Bm
n6=, Im
n6= 0
(5)
with
Ξs=1
Ps,mCqdαq
s,n
.(6)
Bm
n=Y
i∈Bm
n/s
Ξi,(7)
B=Y
i∈B/s
Ξi,(8)
ΘBm
n/s =Y
k∈Bm
n/s ΞkΞs,(9)
and
ΘB−Bm
n/j =Y
l∈B−Bm
n/j ΞlΞj.(10)
Proof. For Bm
n=, we can directly obtain Am
n= 0.
For the case Bm
n6=and with no interference Im
n= 0, we
have
Am
n=P(SI N Rm
n> τ)
=PX
s∈Bm
n
1
Ξs
Hs,n τN0.(11)
In order to obtain the probability density function (PDF) of
Y=P
s∈Bm
n
1
ΞsHs,n, we apply the lemma as follows [35].
Lemma 1. Let (Xi)i=1...n,n2, be independent exponential
random variables with pairwise distinct respective parameters
θi. we have the PDF of their sum as
fX1+X2+...+Xn(X) = n
Y
i=1
θin
X
i=1
eθix
n
Q
k=1,k6=i
(θkθi)
.(12)
Based on Lemma 1, the PDF of Y=Ps∈Bm
n
1
ΞsHs,n is
derived as
fY(y) = Y
i∈Bm
n
ΞiX
s∈Bm
n
eΞsy
Q
k∈Bm
n/s ΞkΞs.(13)
Substituting (9) into (7), we obtain
Am
n=P(SI N Rm
n> τ)
=Z
τN0
fY(y)dy
=Y
i∈Bm
n
ΞiX
s∈Bm
n
eΞsτN0
ΞsQ
k∈Bm
n/s ΞkΞs
=Y
i∈Bm
n/s
ΞiX
s∈Bm
n
eΞsτN0
Q
k∈Bm
n/s ΞkΞs.
(14)
For Bm
n6=φand Im
n6= 0, we employ the change of
variables X=Im
n+N0and Z=Y/X to obtain
Am
n=P(z > τ )
=Z
τ
fZ(z)dz
=Z
τZ
0
xfX(x)fY(xz)dxdz.
(15)
By plugging y=xz into (14), we obtain
fY(xz) = Y
i∈Bm
n
ΞiX
s∈Bm
n
eΞsxz
Q
k∈Bm
n/s ΞkΞs.(16)
Next, we focus on computing fx(x). Employing ti=
li,m
LPmax
i,m Hi,nCmdαq
i,n and t=Im
n, we can rewrite tas
t=XiBBm
n
ti,(17)
with
fti(x)Ξiexp (Ξix),(18)
where Ξi=1
Pi,mCqdαq
i,n
.
Based on Lemma 1, the PDF of P
iBBm
n
tiis derived as
fX(x) = fIm
n(xN0)
=Y
j∈B−Bm
n
ΞjX
j∈B−Bm
n
eΞjN0
Q
l∈B−Bm
n/j
lΞj)eΞjx.
(19)
Combining (15), (16) and (19), we obtain
5
0 2.5 5 7.5 10 12.5 15 17.5 20
0
0.05
0.1
0.15
0.2
0.25
0.3
Transmit Power of Marco BS (W)
Availability with CoMP
Ex.
Sim.
Bn
1={1,2,3,4}
Bn
1={1,2,3}
Bn
1={1,2}
Bn
1={1}
(a)
0 2.5 5 7.5 10 12.5 15 17.5 20
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Transmit Power of Macro BS (W)
Availability with CoMP and SA
Ex.
Sim.
Bn
m={1,2,3,4}
Bn
m={1,2,3}
Bn
m={1,2}
Bn
m={1}
(b)
Fig. 2: (a) Single UE Availability with CoMP, (b) Single UE Availability with CoMP and CA.
Am
n=
Y
i∈B
ΞiX
s∈Bm
nX
j∈B−Bm
n
eΞjN0
Q
k∈Bm
n/s
kΞs)
Z(τ, N0, θj+θsz)
Q
l∈B−Bm
n/j
lΞj),
(20)
where
Z(τ, N0,Ξj+ Ξsz)
=Z
τZ
N0
xejsz)xdxdz
=Z
τN0eN0jsz)
j+ Ξsz)+eN0jsz)
j+ Ξsz)2dz.
(21)
Employing a change of variable of u=N0j+ Ξsz), we
obtain
Z(τ, N0,Ξj+ Ξsz)
=Z
N0jsτ)N0
Ξs
eu
u+N0
Ξs
eu
u2du
=eN0Ξjsτ
ΞsΞj+ Ξsτ.
(22)
Combining (20) and (22), we obtain Am
nwith Bm
n6=φand
Im
n6= 0 as
Am
n=Y
i∈B/s
ΞiX
s∈Bm
nX
j∈B−Bm
n
1
Q
k∈Bm
n/s ΞkΞs
eΞsN0τ
j+ Ξsτ)Q
l∈B−Bm
n/j ΞlΞj.
(23)
B. Availability Validation
To verify the derived analytical results, we plot the an-
alytical curves for the availability with CoMP-JP and the
availability with CoMP-JP and CA using (23) and (4) with
the simulation points using Monte Carlo simulation in Fig.
2a and Fig. 2b, where Ex. denotes the Exact results from (4)
and Sim. denotes the results from Monte Carlo Simulation.
In these two figures, we consider two-tier HetNets, including
single macro BS with Pj,m =46 dBm, 4 pico BSs with Pj,m =
30 dBm, and single UE. The single UE and all these BSs are
randomly deployed in a circle area with radius of 500m, and
the path loss exponent is set as 4. Both figures demonstrate
the well match between the derived analytical results and the
simulation, which proves the accuracy of our derived results.
In Fig. 2a, we plot the availability of the single UE versus
the transmit power at the macro BS with different number
of BSs in the coordinated BS cluster B1
n, with the single
subcarrier in each BS (M= 1). In Fig. 2b, we plot the
availability of the single UE with CA, where the number of
subcarriers at each BS is M= 2 with Cq1= ( 0.375
4π)2, and
Cq2= (0.125
4π)2, respectively. The equal power allocation is
applied in each subcarrier, Note that the CoMP-JP technique
is not used when Bm
n={1}in both figures.
We first observe that the availability of the single UE
increases with increasing the transmit power at the macro
BS and the number of coordinated BSs. However, due to
the uneven distribution of these BSs, the distance between
the UE and BSs are not equal, and the improvement gap
between Bm
n={1,2}and Bm
n={1}is larger than that
between with Bm
n={1,2,3}and Bm
n={1,2}, which
also reveals the importance of UE association for availability
improvement. Comparing Fig. 2a with Fig. 2b, we see that
the availability of the single UE occupying two subcarriers
substantially outperforms that occupying single subcarrier,
which reveals the benefits of CA technique. However, due to
the interference from the non-coordinated BSs, the availability
in Fig. 2b is still far from the availability goal of six nines,
which reveals the importance of using resource allocation to
6
optimize the availability.
III. PROB LE M FORMULATION
The target of this paper is to maximize the minimum
availability among all UEs, which is referred to the Max-
Min Availability optimization problem. To achieve this, an
optimization algorithm is required to perform the optimal UE
association, resource assignment, and power allocation under
the transmit power constraint of each BS. We first present the
optimization problem as
max
v,P min
nAn(24)
s.t. XmPs,m Pmax
s,s∈ B,(24a)
Xn∈N vm
s,n 1,s∈ B,m∈ M,(24b)
Ps,m 0,s∈ B,m∈ M,(24c)
vm
s,n ∈ {0,1},n N ,m∈ M,s∈ B.(24d)
In (24), Anis given by (4). The constraint in (24a) indicates
that the maximum power constraint of each BS, and the
constraints (24b) indicates that each BS can be allocated to at
most one UE. Relying on the strictly increasing characteristic
of logarithmic function [36], we thus convert the optimization
problem of (24) into the following equivalent problem
min
v,P max
nXmln 1Am
n(25)
s.t. XmPs,m Pmax
s,s∈ B,(25a)
Xn∈N vm
s,n 1,s∈ B,m∈ M,(25b)
Ps,m 0,s∈ B,m∈ M,(25c)
vm
s,n ∈ {0,1},n N ,m∈ M,s∈ B.(25d)
Optimizing vm
s,n and Ps,m in (25) is a mixed integer
programming problem, which is generally NP-hard, so the op-
timality for any polynomial time solutions is hard to guarantee.
However, if we relax the integer constraint (vm
s,n ∈ {0,1})by
treating vm
s,n as a sharing factor (0 vm
s,n 1), or determine
vm
s,n by some other method, the optimization problem becomes
a convex optimization problem, because the objective function
is concave and all constraints are linear [37]. Remind that
simply relaxing vm
s,n may results in non-integer solutions of
subcarrier assignment. To avoid that, we propose the two-
step optimization algorithm, where the optimal subcarrier
assignment and UE association achieving (24) with equal
power allocation is determined and fixed first, and then the
optimal power allocation is obtained using Lagrangian dual
based method.
IV. OPTIMIZATION ALGORITHMS
In this section, we propose two optimization algorithms to
solve the max-min optimization problem: the TSOA and the
JTOA. In both algorithms, we divide the resource allocation
into the power allocation, and the subcarrier assignment and
UE association. In TSOA, the optimal subcarrier assignment
and UE association is determined first via heuristic algorithm
under equal power allocation, and the optimal power alloca-
tion is obtained via Lagrangian dual based method with the
predetermined optimal subcarrier assignment. The JTOA is an
integration of the two-step optimization algorithm with the
GA algorithm to achieve the interaction between the first step
and the second step. In this algorithm, the optimal subcarrier
assignment and UE association is obtained by GA, and the
Lagrangian based power allocation is used to evaluate the
fitness of each individual in GA and guide the evolution
process.
A. Two-Step Optimization Algorithm
In this subsection, we present the first step and the second
step of TSOA in the following subsections.
1) Subcarrier assignment and UE association: In this
section, we propose the heuristic algorithm for sub-carrier
assignment to decide {vm
s,n}. The description of Algorithm
1for the subcarrier assignment and UE association is given
in the following.
In the initialization step of Algorithm 1, we assume the
equal power allocation among all BSs and thus the power
allocation vector at any subcarrier of each BS is given as
Ps,m =Pmax
s
M. We denote Fs={1,2,· · · , M }as the avail-
able subcarrier set of the sth BS, and F={F1,F2,· · · ,FS}
as all the available subcarriers of all BSs with the total number
of the available subcarriers as |F| =S×Mduring the
initialization.
After the initialization, the heuristic algorithm performs
three steps: the UE selection,the potential gain calculation
and the subcarrier assignment. At the UE selection step, the
UE with the lowest availability nis selected as the first to be
given access to subcarrier, with the objective of maximizing
the minimum availability among UEs as in (24). If there are
more than one UEs with the same minimum availability, a
random UE among them is selected.
At the potential gain calculation step, we calculate the
potential availability gain before and after single subcarrier m
is allocated to UE n. For an available subcarrier m∈ Fs, we
denote P re(Am
n)and Cur(Am
n)as the availability of UE n
before and after m∈ Fsis allocated to UE n, respectively.
Thus, the potential gain Gm
s,nfor allocating mth subcarrier
of sth BS to nUE is given by
Gm
s,n=Cur(Am
n)P re(Am
n).(26)
Using (26), we calculate and record the potential gains
Gm
s,nfor all the available subcarriers and all the available
BSs. At the subcarrier assignment step, the optimal subcarrier
mof the BS swith the maximum potential gain Gm
s,n
is assigned to the nth UE. Note that only single subcarrier
of a BS is allocated to the nth UE to reduce the possible
impact on the availability of the other UEs. These three steps
are iteratively executed until the available subcarrier set F
becomes empty.
7
Algorithm 1: Heuristic algorithm for subcarrier assign-
ment and UE association
Initialization
Set Am
n= 0,vm
s,n = 0, and Ps,m =Pmax
s
Mfor
n N ,s∈ B,m∈ M
Set F={F1,F2,· · · ,FS}, and Fs={1,2,· · · , M }
while F 6=do
Step1: UE selection
Find UE nsatisfies An< Akk∈ N
Step2: potential gain calculation
for s= 1 to Bdo
for m∈ Fsdo
Calculate and record
Gm
s,n=Cur(Am
n)P re(Am
n)
end
end
Step3: subcarrier assignment
Find a subcarrier min BS ssatisfies
Gm
s,n> Gm
s,nm∈ Fs,s∈ B
Update vm
s,n= 1,Fs=Fs/m, and F=F/m
end
2) Power allocation with fixed subcarrier assignment and
UE association: Fixing the optimized subcarrier assignment
and UE association in the first step, we limit vm
s,n to integer
with fixed value. Thus, we can represent the optimization
problem in (25) as
min
Pmax
nXmln 1Am
n(27)
s.t. XmPs,m Pmax
s,s∈ B,(27a)
Ps,m 0,s∈ B,m∈ M.(27b)
Following [38], we further convert this problem into the
following optimization problem
min
v,p ζ(28)
s.t. XmPs,m Pmax
s,s∈ B,(28a)
Ps,m 0,s∈ B,m∈ M,(28b)
Xmln (1 Am
n)ζ, n N .(28c)
The above optimization problem can be solved by the
Lagrangian dual method, where the corresponding Lagrangian
is written as
L(P;α,β) = ζ+X
s
αsX
m
Ps,m Pmax
s
+X
n
βnX
m
ln(1 Am
n)ζ,
(29)
where α(α0) is the Lagrange multiplier vector associated
with constraint in (28a) and β(β0) is the Lagrange
multiplier vector associated with constraint in (28c). As such,
the dual problem is given by
max
α0,β0min
PL(P;α,β).(30)
The above dual problem can be solved iteratively by decom-
posing it into two nested loops: the inner loop that minimizes
over Pwith the given αand β, and the outer loop that is the
master dual problem maximizing over αand β.
A. Solution to the inner loop: With the given αand β,
minPL(P;α,β)is a standard concave form, and hence we
apply the Karush-Kuhn-Tucker conditions to find the optimal
solution. To solve the inner loop, we first derive the first-order
derivatives as
L
∂Ps,m
=αsX
n
βn
1
1Am
n
∂Am
n
∂Ps,m
,(31)
Note that Am
ndepends on Bm
nand Im
n. If Bm
n=, we can
directly obtain ∂Am
n
∂Ps,m = 0.
If Bm
n6=and Im
n= 0, we derive
∂Am
n
∂Ps,m
=ΩBm
nX
s∈Bm
n
eΞsτN0τN0ΘΘ
Ξs
Θ2
Ξs
∂Ps,m
=Bm
n
P2
s,mCqdαq
s,n X
s∈Bm
n
eΞsτN0
Θ2τN0Θfk,s,
(32)
where
fk,s =X
k∈Bm
n/k,s Y
k∈Bm
n/s ΞkΞs.(33)
If Bm
n6=and Im
n6= 0, we have
∂Am
n
∂Ps,m
=ΞsB
∂Ps,m X
s∈Bm
nX
j∈B−Bm
nΥ
Θ2
Bm
n/s
ΘBm
n/s
Ξs
+Υ
Ξs
1
ΘBm
n/s
=B
P2
s,mCqdαq
s,n X
s∈Bm
nX
j∈B−Bm
n
Υfk,s
Θ2
Bm
n/s
T0(N01)τeθsN0τ
ΘB−Bm
n/j Ξj+ Ξsτ2ΘBm
n/s
,
(34)
where
Υ = eΞsN0τ
Ξj+ ΞsτQ
l∈B−Bm
nΞlΞj.(35)
By substituting (32) and (34) into (31) and setting ∂L
∂Ps,m =
0, we derive the optimal power allocation P
s,m as
8
P
s,m =
0 if |Bm
n|= 0
r=Bm
nP
s∈Bm
n
eΞsτN0
Θ2
Bm
n/s τN0ΘBm
n/s fk,s
if Bm
n6=, Im
n= 0
s=BP
s∈Bm
nP
j∈B−Bm
nΥfk,s
Θ2
Bm
n/s
+(1+N0)τeθsN0τ
ΘB−Bm
n/j (θj+θsτ)2ΘBm
n/s
if Bm
n6=, Im
n6= 0
(36)
where
==1
αsCqdαq
s,n X
n
βn
1Am
n
.(37)
and Bm
n,B,ΘBm
n/j ,ΘB−Bm
n/j ,fk,s and Υare given in (7)
(10), (33) and (35).
Note that the local optimal solution ζand local optimal
availability (Am
n)can be obtained using P
s,m in (36).
B. Solution to the outer loop: To find αand βwith the
obtained Pand ζin the inner loop, we apply the subgradient
method. By doing so, αand βcan be updated iteratively using
αg+1
s="αg
sκPmax
sX
m
P
s,m#+
s∈ B,(38)
βg+1
n="βg
nκζX
m
ln(1 (Am
n))#+
n N ,
(39)
where κ=0.1
gis a diminishing step size, gis the iterative time,
[·]+denotes the updated αand βneeds to be non-negative.
The calculation process of the optimal power allocation
Pand the update process of αand βare repeated until
convergence, where the dual optimal is reached. Knowing that
the inner loop is a convex problem, the duality gap is zero.
The detailed procedures for solving the dual problem in (28)
are illustrated in Algorithm 2.
Algorithm 2: Power allocation based on Lagrangian dual
approach
Set g= 0, and initialize αg,βg, and P
repeat
In prim domain, solve min
PL(P;α,β)to obtain P,
ζand (Am
n)
In dual domain, update the dual variable vector αg+1
according to (38)
In dual domain, update the dual variable vector βg+1
according to (39)
g=g+ 1
until convergence;
In TSOA, Algorithm 2 is executed after Algorithm 1, we
thus calculate the complexity of TSOA as O(M2S2) + O(L),
where Sis the total number of BSs in B, and O(L)is the time
complexity of Algorithm 2, which depends on the complexity
of outer and inner loops. Due to the fact that Algorithm 2
is a standard convex optimization, a fast convergence speed
can be guaranteed for the subgradient method and power
optimization based on KKT condition [20]. We thus can
qualitatively conclude that O(L)is low and acceptable. It also
should be observed that algorithm 2 can only be executed after
determining the subcarrier assignment and UE association,
which means Algorithm 1 and Algorithm 2 can not be
iteratively operated in TSOA. In the next section, we will
design a GA based approach to iteratively update the subcarrier
assignment and UE association, which will finally converge to
the optimal solution. However, much more time complexity is
required.
B. Joint Two-step Optimization based on Genetic Algorithm
In this section, we propose JTOA based on GA to achieve
the interaction between the first step and the second step of
TSOA. GA has already been widely used to tackle many real
world NP-hard problems, such as BS placement optimization
for LTE heterogeneous networks [39] or joint channel and
power allocation for HetNets [40]. This bio-inspired algorith-
m imitates the natural evolution of biological organisms to
provide a robust, near optimal solution for various problems.
In Fig. 3, the JTOA based on GA is illustrated in detail. The
first operation is initialization. Initially, the GA generate Rma-
trices to form the initial population set R={Γr}R
r=1, where
each matrix Γr=γr
s,mS×Mcorresponds to a potential
solution of the UE association and the subcarrier assignment.
Each matrix element γr
s,m Γr(1 rR)denotes that the
γr
s,mth UE is associated with the mth subcarrier of the sth BS
(0 γr
s,m N). Generally, the γr
s,m in the initial population
should be randomly generated to preserve the diversity of the
population. However, considering that the solution obtained
by our proposed heuristic algorithm in Algorithm 1 is also
a sub-optimal solution of the subcarrier assignment and UE
association, we take the sub-optimal solution in Algorithm 1
as a potential individual in the initial population. Thus, the
initial population in our algorithm contains R1randomly
generated individuals and one existing sub-optimal solution.
By doing so, this initialization can converge much faster than
that without using the sub-optimal solution in Algorithm 1,
as validated by the simulation in Fig. 4b.
The initialized Rindividuals only describes the subcarrier
assignment and UE association using R={Γr}R
r=1. We then
perform the power allocation for these Rindividuals using
Algorithm 2 based on the Lagrangian dual method in the
second operation of Fig. 3.
The third operation is the fitness value calculation and the
natural selection. We calculate the fitness values (minimum
UE availability) of all the individuals {Γr}R
r=1 in population
Ras
f(r) = min
nAn,n∈ N (40)
where Anis calculated using (4). Each two individuals are
selected using roulette wheel selection method [41], where
9
Fig. 3: Flowchart describing one iteration of the GA in solving
the optimization problem
the selection probability of each individual is given as
qr=f(r)
Pr∈R f(r).(41)
This selection is repeated until Rindividual are selected.
These selected individuals are used to generate new popula-
tions with crossover and mutation operators.
The fourth operation is the crossover and the mutation. The
conventional two-point crossover is performed to produce new
solutions for two parent individuals in R={Γr}R
r=1. We first
randomly generate two crossover points, then each element
between the two points are switched between two parent
individuals to produce two child individuals. In the mutation
operation, some elements in these two child individuals are
randomly altered to diversify the population and pave the
way towards optima. More specifically, each element of the
matrix in R={Γr}R
r=1 can be mutated or not decided by
the predetermined mutation probability pm. If a element γr
s,m
performs the mutation, a random integer value xbetween 1
and Nwill be chosen to replace the original value (γr
s,m =x).
The sixth operation is the replacement based on an elitist
model, which is used to update a certain number of individuals
in the old population with the new generated individuals. Since
the UE association and subcarrier assignment described by
individuals have been altered during third and fourth operation,
we need to calculate the fitness value of the new generated
individuals using (40) in the fifth operation, then the parent
individuals with the low fitness value will be replaced by
the new generated individuals with higher fitness value in
the next generation in the sixth operation. This population
evolution operations will repeat until convergence, where
the convergence is defined when the maximum fitness value
remains constant for a fixed number of successive iterations
[42].
The JTOA of the UE association and subcarrier assign-
ment, and the power allocation based on GA is described in
Algorithm 3, where Gis the given number of generations,
Ris the population size, pcis the crossover probability,
and pmis the mutation probability. Due to the fact the
Lagrangian dual method is applied to evaluate each individual
in each population, the time complexity of solving this JTOA
is O(GR(O(L) + R)), which showcases the higher imple-
mentation complexity is required to obtain a better solution
compared with that of TSOA.
Algorithm 3: JTOA
set g= 1
Generate initiation population with R1randomly
generated individuals and 1individual by Algorithm 1
Calculate fitness value for each individual in Rusing
Algorithm 2
repeat
for i= 1 to R/2 do
Select two parents p1and p2from Rusing
roulette wheel selection method
r2i1=p1and r2i=p2
Cross r2i1and r2iusing two-point crossover
strategy with probability pc, and produce two
children r0
2i1and r0
2i
Mutate r0
2i1and r0
2iusing mutation strategy
with probability pm
R0=R0r0
2i1, r0
2i
Calculate fitness value for each individual in R0
using Algorithm 2
end
Replace the individuals with low fitness values in
population Rwith the children in offspring R0
until convergence;
Return the fittest individual in R
V. NUMERICAL RESULTS
In this section, we provide numerical results to illustrate the
performance of our proposed algorithm. We consider CoMP
and CA-enabled HetNets consisting of 2 tiers (marco and pico)
with 2 bands (800MHz and 2.5GHz), where each band has a
bandwidth of 10MHz. The set-up is a circle area with size
(π5002) m2, where the macro BS is located at the center, the
pico BSs and UEs are randomly distributed in this circle area.
The details of parameters are summarized in Table I unless
otherwise specified. All the results are obtained by averaging
100 random simulations, and the obtained availability is the
minimum availability among all UEs.
Fig. 4a plots the convergence behavior of our proposed
JTOA with different number of UEs with M= 20, 9 Pico
BSs and 1 Marco BS, where the initial population is a solution
generated by proposed heuristic algorithm in Algorithm 1.
10
0 200 400 600 800 1000 1200 1400 1600 1800 2000
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Number of Generations
Availability
N=4
N=8
N=12
N=16
N=20
0 200 400
0.9
1
(a)
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Number of Generations
Availability
N=15 without heuristic
N=15 with heuristic
N=20 without heuristic
N=20 with heuristic
0 300 600
0.8
0.9
1
(b)
Fig. 4: (a) Convergence behavior with different number of UEs, (b) Convergence comparison with and without heuristic
algorithm.
TABLE I: SIMULATION PARAMETERS
Parameter Value
The number of macro BS 1
The number of pico BS 19
The number of UEs 220
Maximum transmit power of macro BS 46dBm (40W)
Maximum transmit power of pico BS 30dBm (1W)
800MHz band’s wavelength µ10.375m
2.5GHz band’s wavelength µ20.125m
800MHz band’s path loss exponent α13
2.5GHz band’s path loss exponent α24
The number of subcarriers in each band 10
Noise PSD -174dBm
SINR threshold τ1
Population size 20
Crossover probability 0.95
Mutation probability 0.005
Maximum generation 3000
TABLE II: Optimized Availability Value
Tech.
N4 8 12 16 20
CA 10 nines 7 nines 5 nines 3 nines 3 nines
CA & CoMP 15 nines 15 nines 10 nines 6 nines 4 nines
Fig. 4b plots the convergence behavior of JTOA, where the
initial population is a solution generated by proposed heuristic
algorithm in Algorithm 1 or random generated population. In
Fig. 4a, we see that our proposed algorithm converges after
approximately 1000 number of generations for various number
of UEs. However, in Fig. 4b, we notice that our proposed
algorithm converges approximately after 1500 generations if
the initial population is not generated by the sub-optimal
solution in Algorithm 1. This reveals that applying Algorithm
1for initial population generation can speed up convergence.
In both figures, we observe that the converge speed can be
substantially increased with decreasing number of UEs in
HetNets.
In Table II, we also present the optimized availability based
on JTOA for various number of UEs with and without CoMP
in HetNets. We see that with the interference coordination
at each path, the optimized availability is always higher than
that without CoMP, which can be contributed to the single
path gain obtained via CoMP. More importantly, with higher
number of UEs, the benefits of CoMP in achieving high
availability in HetNets become less obvious, which is due to
the reduced number of paths for each UE.
Fig. 5 compares the optimized UE availability based on
our proposed JTOA with that based on TSOA, and the non-
optimized UE availability, and the optima with M= 20,
9 Pico BSs and 1 Marco BS. Here, the optima is obtained
by searching all feasible subcarrier allocations with brute
force approach. Note that Fig. 5a plots the actual availability,
and Fig. 5b plots the corresponding number of nines. We
first observe that the availability decreases with increasing
the number of UEs. This can be explained by the fact that
the transmit power allocated to the UE decreases and the
interference from the same subcarrier at other BSs increases
with increasing the number of UEs. More importantly, the
optimized UE availability based on JTOA outperforms that
based on TSOA, and the non-optimized UE availability, and
closely approaches the performance of the optima obtained
by brute force approach, which showcases the effective of our
proposed JTOA.
Fig. 6 plots the availability versus the number of subcarriers
per BS Mwith 9 Pico BSs and 1 Marco BS, where Fig. 6a
plots the actual availability and Fig. 6b plots the corresponding
number of nines. We can see that the availability increases
with increasing the number of available subcarriers. This can
be explained by the fact that interference decreases with the
increasing the number of subcarriers, thus the single path
availability is improved. This can also be contributed to the
fact that increasing the number subcarriers also increases the
potential gain from the spectral diversity. We find that more
subcarriers are needed to achieve the same availability with
more UEs. In order to achieve a minimum availability of 6
nines, we need at least 15 subcarriers per BS for 15 random
11
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
N
Availability
Non−Optimized
TSOA
JTOA
Optimal
(a)
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
N
Number Of Nine(s)
Non−Optimized
TSOA
JTOA
Optimal
(b)
Fig. 5: Performance comparison with different number of UEs, (a) actual availability, (b) number of nines.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
M
Availability
N=5
N=10
N=15
N=20
(a)
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
M
Number Of Nine(s)
N=5
N=10
N=15
N=20
(b)
Fig. 6: Availability with different number of subcarriers, (a) actual availability, (b) number of nines
123456789
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Number of Pico(s)
Number Of Nine(s)
N=5
N=10
N=15
N=20
Fig. 7: Availability with different number of Pico BSs
deployed UEs.
Fig. 7 plots the availability versus the number of Pico BSs
with M= 20. We can see that the availability increases with
increasing the number of Pico BSs. This is because increasing
the number of Pico BSs increases the number of paths can
be connected for each UE, and decreases the co-channel
interference in the same subcarrier. However, we observe that
availability can not be further increased when the number
of Pico BSs is larger than 8. This indicates that increasing
the number of Pico BSs can not constantly increase the UE
availability.
Fig. 8 plots the availability versus the BS allocated power
ratio ρwith M= 20, 9 Pico BSs and 1 Marco BS, where ρis
the maximum BS allocated power divided by the maximum BS
transmit power. It is shown that the availability increases with
increasing ρ. For the number of UEs is small N=5, 10, or 15,
6 nines availability can be achieved with very low allocated
power ratio ρ= 1/16, whereas for large number of UEs N=
20, 6 nines availability can not be achieved even with full
power allowance ρ= 1. This indicates that increasing the BS
power allocation ratio can not always guarantee substantial
improvement in the availability of HetNets.
12
0 1/8 2/8 3/8 4/8 5/8 6/8 7/8 1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
ρ
Number Of Nine(s)
N=5
N=10
N=15
N=20
Fig. 8: Availability with diferent allocated power ratio
VI. CONCLUSIONS
In this paper, we have presented the theoretical model and
optimization algorithms to achieve high availability in CoMP
and CA-enabled HetNets. We have derived a closed-form
expression for the availability of random UEs in a CoMP
and CA-enabled HetNets. We have formulated an optimization
problems to maximize the minimum availability under the BS
transmit power constraint. To solve the optimization problem,
we have proposed the TSOA, where the subcarrier assignment
and UE association solution is obtained via heuristic algorithm,
and power allocation solution is obtained via Lagrangian dual
based method. Moreover, we have proposed JTOA based on
GA to achieve the interaction between the first step and
the second step. Numerical results show that our proposed
JTOA is effective in achieving ultra-high availability. The high
availability requirement in 5G applications, such as haptic
communications, can be achieved via multiple connectivity
with CA and CoMP. However, when the number of UEs
is small, increasing the number of Pico BSs and the power
allocation ratio can not constantly increase the UE availability.
ACK NOW LE DG ME NT S
This work is supported by the National Natural Sci-
ence Foundation of China under Grant No. 61402096,
No. 61173153 and No. 61572123; the Fundamental Re-
search Funds for the Central Universities under Grant No.
N150404006; the National Science Foundation for Distin-
guished Young Scholars of China under Grant No. 61225012
and No. 71325002; the Specialized Research Fund of the
Doctoral Program of Higher Education for the Priority De-
velopment Areas under Grant No. 20120042130003.
REFERENCES
[1] J. Jia, Y. Deng, S. Ping, H. Aghvami, A. Nallanathan, and X. Wang.
Optimizing availability in CoMP and CA-enabled hetnets. In Proc. IEEE
ICC, pages 1–6, May. 2017.
[2] Gartner. Gartner says 6.4 billion connected ”things” will be
in use in 2016, up 30 percent from 2015. Available at
http://www.gartner.com/newsroom/id/3165317, 2015.
[3] Eckehard Steinbach, Sandra Hirche, Marc Ernst, Fernanda Brandi, Rahul
Chaudhari, Julius Kammerl, and Iason Vittorias. Haptic communication-
s. Proc. IEEE, 100(4):937–956, Apr. 2012.
[4] Manar Jammal, Ali Kanso, and Abdallah Shami. High availability-aware
optimization digest for applications deployment in cloud. In Proc. IEEE
ICC, pages 6822–6828, Jun. 2015.
[5] Zubair Md Fadlullah, Mostafa M Fouda, Nei Kato, Akira Takeuchi,
Noboru Iwasaki, and Yousuke Nozaki. Toward intelligent machine-to-
machine communications in smart grid. IEEE Commun. Mag., 49(4):60–
65, Apr. 2011.
[6] Ge Xiaohu, Cheng Hui, Mao Guoqiang, Yang Yang, and Tu Song.
Vehicular communications for 5g cooperative small-cell networks. IEEE
Trans. Veh. Commun., 65(10):7882–7994, 2016.
[7] Andreas Frotzscher, Ulf Wetzker, Matthias Bauer, Markus Rentschler,
Matthias Beyer, Stefan Elspass, and Henrik Klessig. Requirements and
current solutions of wireless communication in industrial automation. In
Proc. IEEE ICC Workshops, pages 67–72, Jun. 2014.
[8] TR ETSI. 102 889-2 v1. 1.1. Electromagnetic compatibility and Radio
spectrum Matters (ERM) Available at: http://www.etsi.org/technologies-
clusters/technologies/emc.
[9] Gerhard P Fettweis. The tactile internet: applications and challenges.
IEEE Veh. Technol. Mag., 9(1):64–70, Mar. 2014.
[10] Ge Xiaohu, Ye Junliang, Yang Yang, and Li Qiang. User mobility
evaluation for 5g small cell networks based on individual mobility
model. IEEE J. Sel. Areas Commun., 34(3):528–541, 2016.
[11] P Pukite and J Pukite. Modeling for reliability analysis: markov
modeling for reliability, maintainability, safety, and supportability, 1998.
[12] David Ohmann, Meryem Simsek, and Gerhard P Fettweis. Achieving
high availability in wireless networks by an optimal number of rayleigh-
fading links. In Proc. IEEE Globecom Workshops (GC Wkshps), pages
1402–1407, Dec. 2014.
[13] David Ohmann, Willy Teske, and Gerhard P Fettweis. Combining
nakagami-m fading links for high availability in wireless networks. In
Proc. IEEE VTC, pages 1–6, May 2015.
[14] Zhang Rui, Wei Jibo, and Victor Leung. Outage probability of composite
microscopic and macroscopic diversity over correlated shadowed fading
channels. China Communications, 10(11):129–142, Nov. 2013.
[15] Niklas A Johansson, Y-P Eric Wang, Erik Eriksson, and Martin Hessler.
Radio access for ultra-reliable and low-latency 5g communications. In
Proc. IEEE ICC Workshop, pages 1184–1189. IEEE, 2015.
[16] Guillermo Pocovi, Beatriz Soret, Mads Lauridsen, Klaus I Pedersen, and
Preben Mogensen. Signal quality outage analysis for ultra-reliable com-
munications in cellular networks. In Proc. IEEE Glebecom Workshops,
pages 1–6. IEEE, 2015.
[17] Ge Xiaohu, Huang Kun, Wang ChengXiang, Hong Xuemin, and Yang
Xi. Capacity analysis of a multi-cell multi-antenna cooperative cellular
network with co-channel interference. IEEE Trans. Wireless Commun.,
10(10):3298–3309, 2011.
[18] David Oehmann, Ahmad Awada, Ingo Viering, Meryem Simsek, and
Gerhard Fettweis. SINR model with best server association for high
availability studies of wireless networks. IEEE Wireless Commun. Lett.,
5(1):60–63, Feb. 2016.
[19] Haeyoung Lee, Seiamak Vahid, and Klaus Moessner. A survey of radio
resource management for spectrum aggregation in LTE-advanced. IEEE
Commun. Surveys & Tutorials, 16(2):745–760, May 2014.
[20] Guanding Yu, Qimei Chen, Rui Yin, Huazi Zhang, and Geoffrey Ye Li.
Joint downlink and uplink resource allocation for energy-efficient carrier
aggregation. IEEE Trans. Wireless Commun., 14(6):3207–3218, 2015.
[21] J. Jia, Y. Deng, S. Ping, H. Aghvami, and A. Nallanathan. High
availability optimization in heterogeneous cellular networks. In Proc.
IEEE GLOBECOM, pages 1–6, Dec. 2016.
[22] Xingqin Lin, Jeffrey G Andrews, Rapeepat Ratasuk, Bikromadittya
Mondal, and Amitava Ghosh. Carrier aggregation in heterogeneous
cellular networks. In Proc. IEEE ICC, pages 5199–5203, Jun. 2013.
[23] 3GPP. Technical specification group radio access network; coordinated
multi-point operation for LTE physical layer aspects (release 11).
Technical report, Technical Report (36.819), 2011.
[24] Yong Cheng, Marius Pesavento, and Anne Philipp. Joint network
optimization and downlink beamforming for CoMP transmissions us-
ing mixed integer conic programming. IEEE Trans. Signal Process.,
61(16):3972–3987, 2013.
[25] Zhikun Xu, Chenyang Yang, Geoffrey Ye Li, Yalin Liu, and Shugong
Xu. Energy-efficient CoMP precoding in heterogeneous networks. IEEE
Trans. Signal Process., 62(4):1005–1017, 2014.
[26] Mingyi Hong, Ruoyu Sun, Hadi Baligh, and Zhi-Quan Luo. Joint
base station clustering and beamformer design for partial coordinated
13
transmission in heterogeneous networks. IEEE J. Sel. Areas Commun.,
31(2):226–240, 2013.
[27] Ping Xia, Chun-Hung Liu, and Jeffrey G Andrews. Downlink coor-
dinated multi-point with overhead modeling in heterogeneous cellular
networks. IEEE Trans. Wireless Commun., 12(8):4025–4037, 2013.
[28] Ling Liu, Virgile Garcia, Lin Tian, Zhengang Pan, and Jinglin Shi. Joint
clustering and inter-cell resource allocation for CoMP in ultra dense
cellular networks. In Proc. IEEE ICC, pages 2560–2564. IEEE, 2015.
[29] Shiying Han, Boon-Hee Soong, and Quang Duy La. Subcarrier allo-
cation in multi-cell OFDMA wireless networks with non-coherent base
station cooperation and controllable fairness. In Proc. IEEE PIMRC,
pages 524–529. IEEE, 2012.
[30] Xiaoming Wang, Fu-Chun Zheng, Pengcheng Zhu, and Xiaohu You.
Energy-efficient resource allocation in coordinated downlink multicell
ofdma systems. IEEE Trans. Veh. Technol., 65(3):1395–1408, 2016.
[31] Jeffrey G Andrews, Franc¸ois Baccelli, and Radha Krishna Ganti. A
tractable approach to coverage and rate in cellular networks. IEEE Trans.
Commun., 59(11):3122–3134, Nov. 2011.
[32] Harpreet S Dhillon, Radha Krishna Ganti, Francois Baccelli, and Jef-
frey G Andrews. Modeling and analysis of K-tier downlink heteroge-
neous cellular networks. IEEE J. Sel. Areas Commun., 30(3):550–560,
Apr. 2012.
[33] Daewon Lee, Hanbyul Seo, Bruno Clerckx, Eric Hardouin, David
Mazzarese, Satoshi Nagata, and Krishna Sayana. Coordinated multipoint
transmission and reception in lte-advanced: deployment scenarios and
operational challenges. IEEE Commun. Mag., 50(2):148–155, 2012.
[34] Virgile Garcia, Yiqing Zhou, and Jinglin Shi. Coordinated multipoint
transmission in dense cellular networks with user-centric adaptive clus-
tering. IEEE Trans. Wireless Commun., 13(8):4297–4308, 2014.
[35] M´
arton Bal´
azs. Sum of independent exponentials. Available at
http://www. math. bme. hu/balazs/sumexp. pdf, 2005.
[36] K-D Lee and Victor CM Leung. Fair allocation of subcarrier and power
in an OFDMA wireless mesh network. IEEE J. Sel. Areas Commun.,
24(11):2051–2060, 2006.
[37] Dimitris Bertsimas, Omid Nohadani, and Kwong Meng Teo. Nonconvex
robust optimization for problems with constraints. Informs Journal on
Computing, 22(1):44–58, 2010.
[38] Miha Rugelj, Urban Sedlar, Mojca Volk, Janez Sterle, Melita Hajdinjak,
and Andrej Kos. Novel cross-layer qoe-aware radio resource allocation
algorithms in multiuser OFDMA systems. IEEE Trans. Commun.,
62(9):3196–3208, 2014.
[39] Seungseob Lee, SuKyoung Lee, Kyungsoo Kim, and Yoon Hyuk Kim.
Base station placement algorithm for large-scale LTE heterogeneous
networks. PloS one, 10(10):0139190, Oct. 2015.
[40] Kohta Nakamura, Taichi Tashiro, Keiji Yamamoto, and Kenichi Ohno.
Transmit power control and channel assignment for femto cells in hetnet
systems using genetic algorithm. In Proc. IEEE PIMRC, pages 1669–
1674. IEEE, 2015.
[41] David E Goldberg. Genetic algorithms. Pearson Education India, 2006.
[42] Hakim Ghazzai, Elias Yaacoub, Mohamed-Slim Alouini, and Adnan
Abu-Dayya. Optimized smart grid energy procurement for LTE networks
using evolutionary algorithms. IEEE Trans. Veh. Technol., 63(9):4508–
4519, Nov. 2014.
Jie Jia received her PhD degree in Computer sci-
ence and technology from Northeastern University in
2009. She currently works as an associate professor
in Northeastern University, China. In 2016, she
worked as a visiting research associate in King’s
College London (KCL). She is a member of various
international societies such as the IEEE and China
Computer Federation (CCF). She has published over
100 technical papers on various aspects of wireless
networks. Her current research mainly focuses on
HetNets, IoT, and cognitive radio networks.
Yansha Deng (S’13-M’16) received the Ph.D. de-
gree in Electrical Engineering from Queen Mary
University of London, UK, 2015. She is currently
the postdoctoral research fellow in the Department
of Informatics, at King’s College London, UK. Her
research interests include massive MIMO, HetNets,
molecular communication, cognitive radio, coopera-
tive networks, and physical layer security. She has
received the Best Paper Award in ICC 2016. She
is currently an Editor of of IEEE Communications
Letters. She has also served as TPC member for
many IEEE conferences, such as IEEE GLOBECOM and ICC.
Jian Chen received his PhD degree in Computer
science and technology from Northeastern Univer-
sity in 2010. He currently works as an associate
professor in Northeastern University, and as a Senior
Software Engineer in Neusoft Corporation. In 2016,
he worked as a visiting research associate in K-
ing’s College London (KCL). His research interests
include D2D communication, location technology,
network management, signal and image processing.
Abdol-Hamid Aghvami (M’89–SM’91–F’05) is a
Professor of telecommunications engineering at K-
ing’s College London. He joined the academic staff
at King’s College London in 1984. In 1989 he was
promoted to Reader, and in 1993 was promoted
Professor in Telecommunications Engineering. He is
the founder of the Centre for Telecommunications
Research at King’s. He was the Director of the
Centre from 1994 to 2014.
He carries out consulting work on Digital Radio
Communications Systems for British and Interna-
tional companies; he has published over 560 technical journal and conference
papers, filed 30 patents and given invited talks and courses the world over on
various aspects of Mobile Radio Communications. He was Visiting Professor
at NTT Radio Communication Systems Laboratories in 1990, Senior Research
Fellow at BT Laboratories in 1998-1999, and was an Executive Advisor to
Wireless Facilities Inc., USA, in 1996-2002. He is the Chairman of Advanced
Wireless Technology Group Ltd. He is also the Managing Director of Wireless
Multimedia Communications Ltd, his own consultancy company.
He leads an active research team working on numerous mobile and personal
communications projects for Fourth and fifth generation networks; these
projects are supported both by government and industry. He was a member
of the Board of Governors of the IEEE Communications Society in 2001-
2003, was a Distinguished Lecturer of the IEEE Communications Society
in 2004-2007, and has been member, Chairman, and Vice-Chairman of
the technical programme and organising committees of a large number of
international conferences. He is also founder of the International Symposium
on Personal Indoor and Mobile Radio Communications (PIMRC), a major
yearly conference attracting some 1,000 attendees.
He was awarded the IEEE Technical Committee on Personal Communi-
cations (TCPC) Recognition Award in 2005 for his outstanding technical
contributions to the communications field, and for his service to the scientific
and engineering communities. Professor Aghvami is a Fellow of the Royal
Academy of Engineering, Fellow of the IET, Fellow of the IEEE, and in
2009 was awarded a Fellowship of the Wireless World Research Forum
in recognition of his personal contributions to the wireless world, and for
his research achievements as Director at the Centre for Telecommunications
Research at King’s.
14
Arumugam Nallanathan (S’97-M’00-SM’05-F’17)
is Professor of Wireless Communications in the
Department of Informatics at Kings College London
(University of London). He served as the Head
of Graduate Studies in the School of Natural and
Mathematical Sciences at Kings College London,
2011/12. He was an Assistant Professor in the De-
partment of Electrical and Computer Engineering,
National University of Singapore from August 2000
to December 2007. His research interests include
5G Wireless Networks, Internet of Things (IoT) and
Molecular Communications. He published more than 300 technical papers
in scientific journals and international conferences. He is a co-recipient of
the Best Paper Award presented at the IEEE International Conference on
Communications 2016 (ICC2016) and IEEE International Conference on
Ultra-Wideband 2007 (ICUWB 2007). He is an IEEE Distinguished Lecturer.
He has been selected as a Web of Science Highly Cited Researcher in 2016.
He is an Editor for IEEE Transactions on Communications and IEEE Trans-
actions on Vehicular Technology. He was an Editor for IEEE Transactions
on Wireless Communications (2006-2011), IEEE Wireless Communications
Letters and IEEE Signal Processing Letters. He served as the Chair for the
Signal Processing and Communication Electronics Technical Committee of
IEEE Communications Society and Technical Program Chair and member of
Technical Program Committees in numerous IEEE conferences. He received
the IEEE Communications Society SPCE outstanding service award 2012 and
IEEE Communications Society RCC outstanding service award 2014.
... The result revealed that the average reliability is improved by increasing the number of links or the power budget ratio or decreasing the number of UEs. This work was extended by [20], a joint two-step optimization algorithm (JTOA) which is an integration of the two-step optimization algorithm (TSOA) with the GA algorithm to achieve the interaction between the first step and the second step was proposed. ...
... The transmission between the BS and UE is modeled by Rayleigh fading, where H s;n $ expð1Þ. According to [20], A m s;n is given by ...
... Reliability is the goal of optimization in this model, while operation time is also an important indicator. This paper is an extension to [20,21]. The network structure and system model are similar, and the optimization problems are all non-convex. ...
Article
Full-text available
Heterogeneous cellular networks (HetNets) have been proven as a promising approach to deal with ever-growing data traffic. Supporting ultra-reliable and low-latency communication (URLLC) is also considered as a new feature of the upcoming wireless networks. Due to the overlapping structure and the mutual interference between cells in HetNets, existing resource allocation approaches cannot be directly applied for real-time applications, especially for URLLC services. As a novel unsupervised algorithm, Deep Q Network (DQN) has already been applied to many online complex optimization models successfully. However, it may perform badly for resource allocation optimization in HetNets, due to the tiny state change and the large-scale action space characteristics. In order to cope with them, we first propose an auto-encoder to disturb the similarity of adjacent states to enhance the features and then divide the whole decision process into two phases. DQN is applied to solve each phase, respectively, and we iterate the whole process to find the joint optimized solution. We implement our algorithm in 6 scenarios with different numbers of user equipment (UE), redundant links, and sub-carriers. Simulations results demonstrate that our algorithm has good convergence for the optimization objective. Moreover, by further optimizing the power allocation, a 1–2 nines of reliability improvement is obtained for bad conditions. Finally, the experiment result shows that our algorithm reaches the reliability of 8-nines in common scenarios. As an online method, the algorithm proposed in this paper takes only 0.32 s on average.
... In [33], the reliability of C-RAN was studied. In [34], availability analysis and optimization in CoMP and CA-enabled Het-Nets were considered. We are inspired by these recent works. ...
Article
Full-text available
FemtoCaching technology, aiming at maximizing the access probability of streaming media transmission in heterogeneous cellular networks, is investigated in this paper. Firstly, five kinds of streaming media deployment schemes are proposed based on the network topology and the relationship between users and streaming media. Secondly, a matching algorithm for adaptive streaming media deployment is proposed, where the FemtoCaching can be adjusted dynamically. Thirdly, a joint problem is formulated combined with the channel assignment, the power allocation, and the caching deployment. To address this problem, we propose a joint optimization algorithm combining matching algorithm and genetic algorithm to maximize the access probability of streaming media transmission. Simulation experiments demonstrate that: (1) the average access probability of all users accessing streaming media in the network based on the proposed algorithm compared with recent works can be greatly improved, and (2) the performance increases with increasing the number of channels and the storage capacity of micro base stations, but decreases with increasing the number of users.
... Literatures [5]- [8] [9] [12], [16] [14], [15] [13], [17] [18] [19]- [21] Proposed Multi-Numerology CoMP Power Allocation Subcarrier Assignment User Association Throughput Latency Quality-of-Service service, which provides higher flexibility of both frequency and time resources. Additionally, CoMP is capable of mitigating the inter-cell interference (ICI) [22] by transforming interference into desired signals improving system throughput. ...
Preprint
With proliferation of fifth generation (5G) new radio (NR) technology, it is expected to meet the requirement of diverse traffic demands. We have designed a coordinated multi-point (CoMP) enhanced flexible multi-numerology (MN) for 5G-NR networks to improve the network performance in terms of throughput and latency. We have proposed a CoMP enhanced joint subcarrier and power allocation (CESP) scheme which aims at maximizing sum rate under the considerations of transmit power limitation and guaranteed quality-of-service (QoS) including throughput and latency restrictions. By employing difference of two concave functions (D.C.) approximation and abstract Lagrangian duality method, we theoretically transform the original non-convex nonlinear problem into a solvable maximization problem. Moreover, the convergence of our proposed CESP algorithm with D.C. approximation is analytically derived with proofs, and is further validated via numerical results. Simulation results demonstrated that our proposed CESP algorithm outperforms the conventional non-CoMP and single numerology mechanisms along with other existing benchmarks in terms of lower latency and higher throughput under the scenarios of uniform and edge users.
... As specified by 3GPP, SA allows the simultaneous utilization of multiple component carriers (CCs) at physical layer [14]. With SA, we can obtain a large bandwidth of 200MHz by simply aggregating 10 CCs, where each CC has a bandwidth of 20 MHz, thus to improve the bit rates and capacity [15]. There has been some research in applying SA technology for CR networks. ...
Article
Full-text available
As an integration of spectrum aggregation (SA) and cognitive radio Ad Hoc networks (CRAHNs), SA-enabled CRAHNs are capable of utilizing non-continuous spectrum components and is promising in improving the network performance. However, existing research on CRAHNs mainly assuming of an ideal spectrum sensing while ignoring the false alarm probability, resulting in an inaccurate capacity characterization. Hence, in this paper, we propose a more accurate capacity characterization while the detection and false alarm probability are considered. Moreover, we propose the joint optimization model for SA-enabled CRAHNs constrained by QoS requirements of primary users and network resource allocation. We propose the prim-dual method to decompose this problem into two sub-problems: a physical (PHY) layer sub-problem on CC assignment and power allocation, and a network layer sub-problem on route selection. Besides, these two sub-problems are coupled on the link capacity constraint. For sub-problem at PHY layer, we propose the genetic algorithm to obtain the optimal CC assignment and successive convex approximation method to the find the optimal power allocation. For sub-problem at network layer, we propose to apply the standard convex method to find the optimal solution. The numerical simulations demonstrate that network throughput can be improved by increasing the number of CCs. The obtained network throughput outperforms algorithms that without power control, spectrum aggregation or route optimization.
Article
A novel multi-stream carrier aggregation (MSCA) enabled coordinated multi-point (CoMP) framework in heterogeneous cellular networks (HetNets) is proposed. Our objective is to maximize the energy efficiency (EE) of the proposed framework by jointly optimizing base station (BS) clustering, user association, sub-carrier assignment and power allocation. The formulated problem is proved to be NP-hard. By analyzing the structure of the original problem, we decouple it into power allocation sub-problem and sub-carrier assignment sub-problem. For the power allocation sub-problem, we propose sequential minimization technique and Zoutendijk feasible direction method to find the optimal power allocation with fixed sub-carrier assignment. For the sub-carrier assignment sub-problem, we propose genetic algorithm (GA) to represent the BS clustering and sub-carrier assignment, and design two GA based joint optimization frameworks. One is using the power allocation algorithm for fitness evaluation and guiding the evolution of GA. Another is optimizing these two sub-problems in an iterative manner, whereas the optimal sub-carrier assignment is updated timely as any change of the power allocation. Numerical results validate the feasibility, fast convergence, and flexibility of the proposed algorithm, \textcolor{red}{and is superior to conventional MSCA or CoMP enabled HetNets.
Chapter
The enhanced mobile broadband is one of the prime features of 3GPP specified 5G NR technology which defines a maximum broadband speed of 20Gbps. It is to be mentioned that data traffic is expected to grow to the tune of 590% for conventional applications, 770% for mobile video, 950% for mobile virtual reality (VR), and 1320% for M2M/IoT by year 2021 (CISCO Whitepaper, “CISCO Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2016–2021,” Feb. 2017.). To facilitate the need and consumption of high speed data by mobile users for various user applications, mm wave carrier is the only options to be exploited due to their huge available bandwidth. A feasibility study of 60 GHz mm wave link is conducted to be used typically as an access link for 5G eMBB APs in small cell deployment scenario like airports, shopping malls, stadiums, etc. Small cells like femto cells with a radius of 50 m are generally considered for the above mentioned scenarios. A 5G NR link, suitable for femto cell network, is simulated and the performance of the link is evaluated in terms of data throughput and BER in varying conditions of SNR. The results obtained are very promising. The post-decoder BER and throughput at varying received signal power are measured. With carrier aggregation (CA) of five component carriers (CC) and bandwidth 1950.703 MHz (~2 GHz), the combined throughput achieved is 7.17 Gbps. The BER obtained above SNR of 20 dB and received signal power of −60 dBm is almost zero. Our results suggest the potential use of 60 GHz signal as an access link for 5G eMBB APs or could be used as a back haul link in small cell networks.
Article
Designing a protection system for improvement of the availability of connections is an important issue in communication networks under resource constraints. Quantification of availability for networks is a fundamental design element of network-communication services. In this paper, we present a generalized mathematical model for connection availability of an optical communication network with a dedicated protection system. Specifically, the proposed model provides a closed-form relation of connection availability with the path availabilities for the M:N dedicated protection system using probability and combinatorics. We performed various numerical tests for protection systems with different path availabilities to demonstrate the plausibility of the model. To validate the model, Monte-Carlo (MC) simulation was carried out and MC simulation provides a good accordance with the theoretical model, satisfying the theoretical results with the reality. The test results can be used as a benchmark for any new model validation in future. The same model can be applied to similar problems related to probabilistic analysis in other areas.
Article
This article investigates the ultra-reliable communication for industrial wireless sensor networks (IWSNs), where frequency diversity, path diversity and cooperative diversity are jointly investigated. We propose a novel IWSN equipped with the full-duplex (FD) relays and carrier aggregation (CA) technique, to improve the reliability of all sensors. We derive a closed-form expression for the reliability characterization based on signal-to-interference-plus-noise (SINR) model for sensors connecting to sink node either via FD relays or direct communications. We formulate a joint resource allocation problem for reliability maximization, considering sub-carrier assignment, relay selection and power control. We propose to apply the distributed decision making (DDM) framework to decouple the problem into two sub-problems and find the joint optimal solution in an iterative manner. We further propose an improved artificial bee colony algorithm to obtain the optimal results for each sub-problem. Our simulation results showcase the reliability increases with increasing the number of FD-relays and sub-carriers. We also show that the FD and CA techniques along with our proposed algorithm can be an effective way for improving the reliability of IWSNs.
Article
Full-text available
Data traffic demands in cellular networks today are increasing at an exponential rate, giving rise to the development of heterogeneous networks (HetNets), in which small cells complement traditional macro cells by extending coverage to indoor areas. However, the deployment of small cells as parts of HetNets creates a key challenge for operators’ careful network planning. In particular, massive and unplanned deployment of base stations can cause high interference, resulting in highly degrading network performance. Although different mathematical modeling and optimization methods have been used to approach various problems related to this issue, most traditional network planning models are ill-equipped to deal with HetNet-specific characteristics due to their focus on classical cellular network designs. Furthermore, increased wireless data demands have driven mobile operators to roll out large-scale networks of small long term evolution (LTE) cells. Therefore, in this paper, we aim to derive an optimum network planning algorithm for large-scale LTE HetNets. Recently, attempts have been made to apply evolutionary algorithms (EAs) to the field of radio network planning, since they are characterized as global optimization methods. Yet, EA performance often deteriorates rapidly with the growth of search space dimensionality. To overcome this limitation when designing optimum network deployments for large-scale LTE HetNets, we attempt to decompose the problem and tackle its subcomponents individually. Particularly noting that some HetNet cells have strong correlations due to inter-cell interference, we propose a correlation grouping approach in which cells are grouped together according to their mutual interference. Both the simulation and analytical results indicate that the proposed solution outperforms the random-grouping based EA as well as an EA that detects interacting variables by monitoring the changes in the objective function algorithm in terms of system throughput performance.
Article
The cooperative transmission is an effective approach for vehicular communications to improve the wireless transmission capacity and reliability in the fifth generation (5G) small cell networks. Based on distances between the vehicle and cooperative small cell BSs, the cooperative probability and the coverage probability have been derived for 5G cooperative small cell networks where small cell base stations (BSs) follow Poisson point process distributions. Furthermore, the vehicular handoff rate and the vehicular overhead ratio have been proposed to evaluate the vehicular mobility performance in 5G cooperative small cell networks. To balance the vehicular communication capacity and the vehicular handoff ratio, an optimal vehicular overhead ratio can be achieved by adjusting the cooperative threshold of 5G cooperative small cell networks.
Article
The advanced metering infrastructure of the smart grid presents the biggest growth potential in the machine-to-machine market today. Spurred by recent advances in M2M technologies, SG smart meters are expected not to require human intervention in characterizing power requirements and energy distribution. However, there are many challenges in the design of the SG communications network whereby the electrical appliances and smart meters are able to exchange information pertaining to varying power requirements. Furthermore, different types of M2M gateways are required at different points (e.g., at home, in the building, at the neighborhood, and so forth) of the SG communication network. This article surveys a number of existing communication technologies that can be adopted for M2M communication in SG. Among these, the most reliable technology to facilitate M2M communication in the SG home area network is pointed out, and its shortcoming is also noted. Furthermore, a possible solution to deal with this shortcoming to improve SG communications scalability is also presented.
Article
With small cell networks becoming core parts of the fifth generation (5G) cellular networks, it is an important problem to evaluate the impact of user mobility on 5G small cell networks. However, the tendency and clustering habits in human activities have not been considered in traditional user mobility models. In this paper, human tendency and clustering behaviors are first considered to evaluate the user mobility performance for 5G small cell networks based on individual mobility model (IMM). As key contributions, user pause probability, user arrival and departure probabilities are derived in this paper for evaluat-ing the user mobility performance in a hotspot-type 5G small cell network. Furthermore, coverage probabilities of small cell and macro cell BSs are derived for all users in 5G small cell networks, respectively. Compared with the traditional random waypoint (RWP) model, IMM provides a different viewpoint to investigate the impact of human tendency and clustering behaviors on the performance of 5G small cell networks.
Article
Fifth generation wireless networks are currently being developed to handle a wide range of new use cases. One important emerging area is ultra-reliable communication with guaranteed low latencies well beyond what current wireless technologies can provide. In this paper, we explore the viability of using wireless communication for low-latency, high-reliability communication in an example scenario of factory automation, and outline important design choices for such a system. We show that it is possible to achieve very low error rates and latencies over a radio channel, also when considering fast fading signal and interference, channel estimation errors, and antenna correlation. The most important tool to ensure high reliability is diversity, and low latency is achieved by using short transmission intervals without retransmissions, which, however, introduces a natural restriction on coverage area.