Content uploaded by Mohammad Reza Zamani
Author content
All content in this area was uploaded by Mohammad Reza Zamani on Apr 13, 2020
Content may be subject to copyright.
Content uploaded by Mohammad Reza Zamani
Author content
All content in this area was uploaded by Mohammad Reza Zamani on Mar 25, 2020
Content may be subject to copyright.
Content uploaded by Mohammad Reza Zamani
Author content
All content in this area was uploaded by Mohammad Reza Zamani on Mar 25, 2020
Content may be subject to copyright.
Joint Successive Base Station Switch Off and User
Subcarrier Allocation Optimization for Green
Multicarrier based Cellular Networks
Danial Davarpanah, Mohammadreza Zamani, Mohsen Eslami, and Taher Niknam
Department of Electrical and Electronics Engineering
Shiraz University of Technology
Shiraz - Iran
Email: {d.davarpanah, m.eslami, zamani, niknam}@sutech.ac.ir
Abstract—In recent years, energy consumption in cellular
networks has dramatically increased. That is mainly due to the
increase of network users and accordingly base stations (BSs).
In this paper we focus on the main component of wireless
infrastructure, which is the BS, and the major energy consumers
in the cellular network. We propose a novel approach using
Teaching-learning based optimization (TLBO) for the energy-
aware management of radio access networks and determine the
optimum operational mode of each BS with respect to energy
efficiency. The proposed algorithm reduces the number of active
BSs when they are underutilized. When some BSs are switched
off, radio coverage and services are taken care of by the cells
that remain active, to ensure user coverage. We demonstrate
via analysis and simulations that significant reduction in energy
consumption is achieved by dynamically switching off some BSs,
While maintaining the Quality of Service (QoS) for all users.
I. INTRODUCTION
Serious impacts of greenhouse effects include substantial
increase in the atmospheric concentration of carbon dioxide
(CO2) [1]. The scope of reducing greenhouse gasses include
many different industry areas, especially the ones more subject
of human daily lives, such as information and communications
technology (ICT). Cisco [2] has predicted that cellular data
traffic will have an annual 78% increase between 2011 and
2016, leading to an overall 18 fold increase. Therefore, it is not
difficult to imagine that consistent with astonishing increase
of cellular data traffic and internet access requests via smart
phones and other portable wireless devices, the number of
cellular base stations to support such amount of data traffic
and increase network capacity and spectral efficiency should
increase. Currently the number of base stations is growing
at an astonishing rate. It is estimated that each year around
120,000 new base stations are deployed to provide service
to around 400 million users worldwide [3]. Hence, keeping
up with information capacity demands and at the same time
avoiding high energy bills has become a serious challenge for
cellular operators.
For all network operators and university/industry research
centers, energy efficiency has become one of the main con-
cerns in designing cellular networks [4]. The main target
of commercial cellular network designers has traditionally
been increasing capacity, data rate, diversity, robustness of
services, and ubiquitous access. Nonetheless, energy efficiency
of cellular networks has now become a decisive criterion.
All BSs have dynamic and constant electrical energy us-
age. The BS dynamic power is used for transmitting signal
to users. The constant part is used for operations such as
signal processing, cooling, power supply, and backup battery
charging, which are independent from traffic load [5]. During
the operation time of a BS, even when there is no traffic in
the system, the BS consumes considerable amount of energy.
Recently compelling amount of research on reducing power
consumption of fixed and mobile wireless systems has
emerged [6]–[8]. Among these works, some only consider the
dynamic power consumption of BSs [9]–[12]. Some others
have taken BS’s fixed power consumption as well. [13],
[14] have introduced ”cell zooming”, in which BSs adjust
their coverage area according to network’s circumstances or
traffic in order to achieve traffic balance between cells while
reducing network’s power consumption. An extreme case of
cell zooming occurs when a cell’s coverage area approaches
zero, or in other words the BS is switched off. [15] has
considered practical and implementation aspects of BS on/off
switching, such as activation time, and ping-pong effect, and
their effects on user QoS and system stability during switching
period.
The aim of this article is to present an energy efficient
subcarrier allocation and BS switch off algorithm in down-
link of an OFDM based cellular network. By dynamically
switching BSs with low traffic off, their traffic is transferred to
neighboring active BSs. However, there are two concerns that
need to be addressed. One is that the neighboring BSs should
have enough resources to accommodate new users. Second,
the total power of the network should not increase due to
the larger distances of the remaining BSs to the users, which
results in larger path loss and increased dynamic power of
the remaining active BSs. Simultaneous search for BS candi-
dates to be switched off and assigning unoccupied subcarriers
of remaining active BSs to their users is a complex mix-
integer and nonlinear optimization problem. Therefore, robust
and reliable optimization techniques are required to solve
2015 23rd Iranian Conference on Electrical Engineering (ICEE)
978-1-4799-1972-7/15/$31.00 c
2015 IEEE 504
the aforementioned problem. In this work, Teaching-learning
based optimization (TLBO) is adopted to find the best power
efficient solution. The proposed algorithm is shown to be very
effective in reducing network’s total power consumption, while
satisfying guaranteed user QoSs.
The remainder of the paper is organized as follows: System
model and problem formulation are presented in Section II.
The proposed BS switch off algorithm is presented in Section
III followed by simulations results in Section IV. Finally,
Section V concludes the paper.
II. SYSTEM MODEL AND PROBLEM FORMULATION
Let us consider a conventional cellular network with a
central cell surrounded by six uniform hexagonal cells. Base
stations are located at the center of the cells being connected
through backbone high speed fiber to a central processing
unit that allows them to exchange user link and traffic in-
formation. For the wireless channel, distance dependent path
loss, log-normal shadowing and multipath Rayleigh fading are
considered. Then the total propagation loss L(dB) at distance
d(m) is represented as: L(dB) = L0+10βlog10 d+XdB,
where L0=31.5dB is the path loss at d=1m and β
is the path loss exponent; shadowing, XdB, is assumed to
be a Gaussian random variable with zero mean and variance
σ2
XdB. Regarding the BS power consumption in a cell, we
consider the average power consumption of a BS consists of
a dynamic and a constant part.
We are searching for BSs that once they are switched off,
network’s total power consumption is reduced. The problem
can be formulated as an optimization problem as follows:
min B
b=1
U
u=1
NC
nc=1
γ(b,u,nc)α(b,u,nc)+ξ
B
b=1
ρ(b),(1)
subject to,
U
u=1
α(b,u,nc)≤1,b=1, ..., B u =1, ..., U
B
b=1
α(b,u)=1,u=1, ..., U
B
b=1
NC
nc=1
α(b,u,nc)r(b,u,nc)≥q(u),u=1, ..., U
(2)
where PTis the network’s total power consumption and the
following binary decision variables are introduced:
α(b,u,nc)=⎧
⎨
⎩
1,if channel ncof BS b
is assigned to UE u
0,otherwise
(3)
which indicates whether channel ncof BS bis assigned to UE
u,
α(b,u)=⎧
⎪
⎨
⎪
⎩
1,if
NC
nc=1
α(b,u,nc)>0
0,otherwise
(4)
which indicates whether BS bserves UE u, and
ρ(b)=⎧
⎪
⎨
⎪
⎩
1,if
U
u=1
α(b,u)>0
0,otherwise
(5)
which indicates whether a BS is ON and serves some UEs or
OFF and serves no UE.
III. THE PROPOSED ALGORITHM
To solve the complex optimization problem of (1), one
needs to find a suitable solution in the mature discipline
of mathematical optimization. There many nature-inspired
optimization algorithms [16]–[19], such as genetic algorithm
(GA), particle swarm optimization (PSO), artificial bee colony
(ABC), ant colony optimization (ACO), teaching-learning
based optimization (TLBO) algorithm. The aforementioned
algorithms have been used for many engineering problems and
proven to be efficient in some cases. Most of these algorithm
require tuning algorithm-specific parameters with special care.
Among these algorithms, TLBO [20] has proven to be
an efficient optimization method for large scale non linear
optimization problems in finding global solutions. As opposed
to other aforesaid algorithms, TLBO does not require tuning
any parameter, which along with high convergence rate make
it a superior candidate compared to its counterparts.
A. Teaching-Learning Based Optimization (TLBO) Algorithm
The TLBO algorithm is a new efficient optimization algo-
rithm which has been inspired by learning mechanism in a
class.The TLBO method is based on the effect of the influence
of a teacher on the output of learners in a class. Here, output
is considered in terms of results or grades.The population is
considered as a group of learners or a class of learners. The
teacher is generally considered as a highly learned person
who shares his or her knowledge with the learners. The two
elementary components of this algorithm are Teacher and
Learners. Based on two basic modes of the learning, through
teacher (known as teacher phase) and interacting with the other
learners (known as learner phase), the procedure of TLBO is
divided into two parts, 1) Teacher phase and 2) Learner phase.
In the teacher phase, the teacher improves the knowledge
of the learners up to the level of his/her own knowledge level.
In fact, in this phase, the quality of the learners is affected by
the good quality of the teacher as the best individual and the
Learner Phase means learning through the interaction between
learners.
To solve (1) with constraints given in (2), The proposed
algorithm works as follows:
i) Traffic load of each BS in terms of required number of
OFDM subchannels is evaluated. Number of required subchan-
nels is obtained based on the capacity of each subchannel and
each user’s requested QoS in terms of bit rate, ii) Network’s
total power consumption, P(i)
T, evaluated and the BS with
minimum traffic load is considered as the candidate to be
switched off, iii) A list of unoccupied subchannels for each
2015 23rd Iranian Conference on Electrical Engineering (ICEE)
505
TABLE I
SIMULATION PARAMETERS
Parameter Value
Constant Power of BS 600 W
Cellular layout hexagonal grid 3-sector sites
Cell Radius 1Km
Frequency reuse 1
Number of Cell-sites 7
BS Maximal TX Power 40 W
Path loss Exponent 2.5
Shadowing Standard Deviation 8dB
System Bandwidth 20 MHz
User Distribution Uniform
Subcarrier Spacing 200 kHz
base station to be used for accommodating users of the
switched off base station is prepared. TLBO algorithm is
executed and the why of reassignment subchannel to users for
minimizing network total power consumption is determined,
and iv) If there is no QoS degradation and the total power of
the network is reduced, the candidate BS is switched off and
the algorithm returns to step(i). The algorithm continues with
the next candidate BS, until the maximum number of BSs is
switched off. Further details of the algorithm are presented in
the journal version of this paper.
IV. SIMULATION RESULTS
To evaluate the performance of the proposed scheme a net-
work of 7macro BSs as described in section II is considered.
The system parameters are summarized in Table I.
Figures 1 and 2 show network’s total power consumption
at each stage of successive BS switching off. The results have
been obtained by averaging network’s total power consump-
tion, PT, for 50 time runs over a fixed network realization
(fixed user channels).
0 1 2 3 4 5 6
1500
2000
2500
3000
3500
4000
4500
Algorithm step index
Total power consumption [W]
Proposed Algorithm
Random Switch
Fig. 1. Comparison of energy saving algorithms
As Figure 1 shows, by switching off each BS, the total
power consumption of the network is reduced. For the 6th
0 1 2 3 4 5 6
1000
1500
2000
2500
3000
3500
4000
4500
Algorithm step index
Total power consumption [W]
Proposed Algorithm
Random Switch
Fig. 2. Comparison of energy saving algorithms
BS, the power reduction is incremental as almost the same
amount of power that is saved from eliminating Pconst of
switched off BSs, needs to be added to Pdynamic of active
BSs to compensate the distance-based channel path loss.
This power trade-off can be better seen in Figure 2 where
network’s average total power increases after switching off
the sixth BS. For comparison, randomly switching BSs off
one at a time has been compared with the proposed scheme.
Proposed algorithm for its powerful optimization core is able
to better allocate subchannel to users and reduce network’s
power consumption.
V. C ONCLUSIONS
An energy efficient solution to for dynamic operation of
mobile networks was presented. Identical BSs were consid-
ered, and the optimal number of active BSs for which user
QoSs were satisfied and minimum total power required was
proposed. Numerical results indicated that proposed scheme
is able to significantly reduce network’s power consumption.
Comparing the obtained values of proposed scheme with that
of random BS switch off showed a significant advantage of
the proposed scheme over its counterpart.
REFERENCES
[1] G. Fettweis and E. Zimmermann, “Ict energy consumption-trends and
challenges,” in Proceedings of the 11th International Symposium on
Wireless Personal Multimedia Communications, vol. 2, no. 4, 2008, p. 6.
[2] Cisco, “Cisco visual networking index: Global mobile data traffic
forecast update, 2011c2016,” Cisco White Paper, Feb. 2012.
[3] H. Sistek, “Green-tech base stations cut diesel usage by 80 percent,”
CNET News Green Tech, 2008.
[4] E. Hossain, V. K. Bhargava, and G. P. Fettweis, Green radio communi-
cation networks. Cambridge University Press, 2012.
[5] O. Arnold, F. Richter, G. Fettweis, and O. Blume, “Power consumption
modeling of different base station types in heterogeneous cellular
networks,” in Future Network and Mobile Summit, 2010. IEEE, 2010,
pp. 1–8.
[6] Z. Hasan, H. Boostanimehr, and V. K. Bhargava, “Green cellular net-
works: A survey, some research issues and challenges,” Communications
Surveys & Tutorials, IEEE, vol. 13, no. 4, pp. 524–540, 2011.
2015 23rd Iranian Conference on Electrical Engineering (ICEE)
506
[7] R. Bolla, R. Bruschi, F. Davoli, and F. Cucchietti, “Energy efficiency
in the future internet: a survey of existing approaches and trends in
energy-aware fixed network infrastructures,” Communications Surveys
& Tutorials, IEEE, vol. 13, no. 2, pp. 223–244, 2011.
[8] A. Amanna, “Green communications: Annotated review and research
vision,” Virginia Tech, p. 87, 2010.
[9] B. Badic, T. O’Farrrell, P. Loskot, and J. He, “Energy efficient radio
access architectures for green radio: large versus small cell size deploy-
ment,” in Vehicular Technology Conference Fall (VTC 2009-Fall), 2009
IEEE 70th. IEEE, 2009, pp. 1–5.
[10] T. Le and M. Nakhai, “Possible power-saving gains by dividing a cell
into tiers of smaller cells,” Electronics letters, vol. 46, no. 16, pp. 1163–
1165, 2010.
[11] B. Badic, T. O’Farrell, P. Loskot, and J. He, “Effect of the base station
antenna beam tilting on energy consumption in cellular networks,” in
Vehicular Technology Conference Fall (VTC 2010-Fall), 2010 IEEE
72nd. IEEE, 2010, pp. 1–5.
[12] F. Cao and Z. Fan, “The tradeoff between energy efficiency and system
performance of femtocell deployment,” in Wireless Communication
Systems (ISWCS), 2010 7th International Symposium on. IEEE, 2010,
pp. 315–319.
[13] Z. Niu, Y. Wu, J. Gong, and Z. Yang, “Cell zooming for cost-efficient
green cellular networks,” Communications Magazine, IEEE, vol. 48,
no. 11, pp. 74–79, 2010.
[14] S. Bhaumik, G. Narlikar, S. Chattopadhyay, and S. Kanugovi, “Breathe
to stay cool: adjusting cell sizes to reduce energy consumption,” in Pro-
ceedings of the first ACM SIGCOMM workshop on Green networking.
ACM, 2010, pp. 41–46.
[15] S.-E. Elayoubi, L. Saker, and T. Chahed, “Optimal control for base
station sleep mode in energy efficient radio access networks,” in INFO-
COM, 2011 Proceedings IEEE. IEEE, 2011, pp. 106–110.
[16] D. E. Goldberg, “Genetic algorithms in search, optimization and ma-
chine learning,” New York: Addison-Wesley, 1989.
[17] M. Clerc, Particle swarm optimization. ISTE Publishing Company,
2006.
[18] D. Karaboga, “An idea based on honey bee swarm for numerical opti-
mization,” Technical report-tr06, Erciyes university, engineering faculty,
computer engineering department, Tech. Rep., 2005.
[19] M. Dorigo and T. Stutzle, Ant colony optimization. MIT Press, 2004.
[20] R. Rao, V. Savsani, and D. Vakharia, “Teaching–learning-based opti-
mization: A novel method for constrained mechanical design optimiza-
tion problems,” Computer-Aided Design, vol. 43, no. 3, pp. 303–315,
2011.
2015 23rd Iranian Conference on Electrical Engineering (ICEE)
507
