Available via license: CC BY-NC-ND 4.0

Content may be subject to copyright.

ScienceDirect

Available online at www.sciencedirect.com

Procedia Computer Science 131 (2018) 1039–1047

1877-0509 © 2018 The Authors. Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

Selection and peer-review under responsibility of the scientiﬁc committee of the 8th International Congress of Information and

Communication Technology

10.1016/j.procs.2018.04.255

10.1016/j.procs.2018.04.255

© 2018 The Authors. Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

Selection and peer-review under responsibility of the scientic committee of the 8th International Congress of Information and

Communication Technology.

1877-0509

Available online at www.sciencedirect.com

ScienceDirect

Procedia Computer Science00 (2018) 000–000

www.elsevier.com/locate/procedia

2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license

https://creativecommons.org/licenses/by-nc-nd/4.0/)

Selection and peer-review under responsibility of the scientific committee of the 8th International Congress of Information and Communication

Technology

8th International Congress of Information and Communication Technology (ICICT-2018)

Research of Resource Allocation Technology Based on MIMO

Ultra Density Heterogeneous Network for 5G

Xie Chaochena

一

, Tan Xiaohenga,, Turdybek Balginbeka, Qu Qiana, Zhang Xiaolianga

aCollege of Communication Engineering, Chongqing University, Chongqing 400044 china

Abstract

With wireless device extensively application and continuing to enlarge accessing network, energy and spectrum resources are

required more widely. Therefore, the purpose of this paper is to improve the resources, spectrum and energy efficiency of the

ultra-density heterogeneous cellular network by using existing technologies, such as enhancing frequency spectrum utilization

ratio orthogonal frequency division multiplexing (OFDM), and combatting channel fading performance MIMO etc. This paper

proposes a method to decrease the whole system power consumption by deploying low power cellular with multi antennas,

max system capacity and coverage. Simulation results show that the method can decrease system power consumption, enhance

throughput and improve system performance. Therefore, multi- antenna ultra-density network resource allocation scheme is

effective.

Keywords: energy efficiency, ultra-density network, heterogeneous, OFDM, MIMO

1. Introduction

Recently, the fifth generation mobile communication technology (5G) is promoted significantly because of

increasing new traffic types and data services (such as the smart grid, smart homes and cities, and e-health etc.).

Thus a new era of information society will be coming: in any time and any place, devices connect to internet will

obtain personalized information service. “Things net” will be turned into reality. The number of devices

accessing network will surpass the population on earth in 2020, so it will consume more energy. Thus a green,

energy saving and environmental protection is an important issue in 5G network. Resource management is an

important aspect to wireless communication system. Owning to 5G network is an extremely problem, which

includes different accessing networks. As a result, resource management confronts huge challenges. Work[1]

analyzed energy consumption problems of different communication networks. Although reducing the transmission

power of the base station saves total power consumption, the design and optimization of network performance

一* Corresponding author. Tel.:18306082918.

E-mail address:xie_cc1@163.com

Available online at www.sciencedirect.com

ScienceDirect

Procedia Computer Science00 (2018) 000–000

www.elsevier.com/locate/procedia

2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license

https://creativecommons.org/licenses/by-nc-nd/4.0/)

Selection and peer-review under responsibility of the scientific committee of the 8th International Congress of Information and Communication

Technology

8th International Congress of Information and Communication Technology (ICICT-2018)

Research of Resource Allocation Technology Based on MIMO

Ultra Density Heterogeneous Network for 5G

Xie Chaochena

一

, Tan Xiaohenga,, Turdybek Balginbeka, Qu Qiana, Zhang Xiaolianga

aCollege of Communication Engineering, Chongqing University, Chongqing 400044 china

Abstract

With wireless device extensively application and continuing to enlarge accessing network, energy and spectrum resources are

required more widely. Therefore, the purpose of this paper is to improve the resources, spectrum and energy efficiency of the

ultra-density heterogeneous cellular network by using existing technologies, such as enhancing frequency spectrum utilization

ratio orthogonal frequency division multiplexing (OFDM), and combatting channel fading performance MIMO etc. This paper

proposes a method to decrease the whole system power consumption by deploying low power cellular with multi antennas,

max system capacity and coverage. Simulation results show that the method can decrease system power consumption, enhance

throughput and improve system performance. Therefore, multi- antenna ultra-density network resource allocation scheme is

effective.

Keywords: energy efficiency, ultra-density network, heterogeneous, OFDM, MIMO

1. Introduction

Recently, the fifth generation mobile communication technology (5G) is promoted significantly because of

increasing new traffic types and data services (such as the smart grid, smart homes and cities, and e-health etc.).

Thus a new era of information society will be coming: in any time and any place, devices connect to internet will

obtain personalized information service. “Things net” will be turned into reality. The number of devices

accessing network will surpass the population on earth in 2020, so it will consume more energy. Thus a green,

energy saving and environmental protection is an important issue in 5G network. Resource management is an

important aspect to wireless communication system. Owning to 5G network is an extremely problem, which

includes different accessing networks. As a result, resource management confronts huge challenges. Work[1]

analyzed energy consumption problems of different communication networks. Although reducing the transmission

power of the base station saves total power consumption, the design and optimization of network performance

一* Corresponding author. Tel.:18306082918.

E-mail address:xie_cc1@163.com

1040 Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047

reduce the energy consumption by in average 80%2. Emmett Carolan et al. chose an ANN employing Back

Propagation (BP) to minimize power consumption in the cellular and the base station3. In addition, it is promising

the solution of enhancing energy efficiency to recourse management4. Energy efficient problem has been

investigated in [5-8]. Work[9] stated energy efficient allocation algorithms based on quality of server (QoS)

constrain. Although extensive studies have researched on these issues, it still lacks enough analysis and researches

on multiple antenna ultra-dense heterogeneous networks. This paper analyzes that energy efficiency, capacity and

covering problem based on 5G multiple antenna ultra-density networks. This paper is organized as follows:

section 2 introduces the architecture of networks and some attribute of networks such as multi mechanisms,

dynamic channel allocation and freely accessing. Section 3 states system model and simulation. Section 4 and 5

analyze the results and make a summary.

2．Network architecture

The 5G application environments and the network Framework are shown in fig.1 and fig.2. Obviously, the 5G

network is very complex, which is combined with the macro base station, the micro base station, the pic-cellular,

the millimeter wave (mWave) and so on. The organization form of heterogeneous and wireless self-organization

net (SON) are adopted. Therefore, this network should be flexible enough to adapt to various scenarios. It

includes 2G base station, 3G base station, 4G and LTE-A, and provides open wireless network accessing.

Fig.1. The deployment environment of the 5G7Fig.2. 5G network architecture Fig.3. Ultra density network for 5G

Fig.2 indicates 5G network architecture. The open wireless networking (OWN) combined with present cloud

technique provides more flexible for radio access. SON is a core of the whole network management. Therefore,

this architecture is called multi-mechanisms cloud accessing network (MMCLAN).

According to above description, 5G network has the following attributes:

2.1 Multi Communication Mechanisms Coexist

For satisfying the requirement of mobility, covering and capacity, solving ubiquitous and wireless accessing,

the 5G network integrates multi-communication mechanism. And it includes a variety of communication systems.

2.2 Intelligent Resource Management

Because of the complexion of network architecture, resource management confronts huge challenge. Intelligent

is a developing direction of network resource management. Increasing spectrum efficiency, saving power

consumption, improving quality of service and optimizing network need intelligent network management

technique. Therefore, 5G network provides more flexible and intelligent recourse allocation and management than

the present network.

Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047 1041

Chaochen Xie/ Procedia Computer Science00 (2018) 000–000

2.3 Open Accessing

5G network is open wireless network10. As various new type technologies and applications emerge constantly,

all of them are included in 5G network. Hence only open wireless radio accessing interface can be satisfied with a

variety of applications freely access to communication system. And the 5G network will be more open and more

flexible.

3. Resource management system model

The system model and assumption are described as following:

3.1 Deployment Scenario

Base stations are deployed in the urban hot spot. An area S is completely covered by various types of cellular

base stations. Fig.3 shows base station deployment. 5G base station owning omnidirectional antenna plays as the

central controller lying in the center of the area, which is responsible for the nodes and resource allocation of the

whole network system. The macro cellular consists of N cellular base station. We define N: {1, 2,…,N} as the set

of cellular network base stations. Each of base station has l transmitting antennas.

3.2 Power Model

The total power consumption of the BS is divided into two parts8. (1) The power consumed in the BS when

there is no transmission data; (2) Power consumption when data are transmitted. Base station consumes power can

be written

0tx

P P P

(1)

Where Ptx denotes transmission power when data are transmitted. Factor

denotes efficiency factor under

various fading conditions such as large scale fading model and small scale fading, shadowing fading and

frequency selective fading etc. P0 is inherently loss such as the active site cooling and signal processing. The base

station transmitting power model given by [7] can be generalized to all BS types, such as macro, micro, pico and

femto BSs. Power consumption for different types of BSs is shown in the table 1.

Table 1: Power Consumption Parameters for Base Station7

Base Station Type Pmax [W] α P0[W

]

Macro 40

2.6

6118.7

Micro 6.3 3.1 53

Pico 0.13 4.2 6.8

3.3 System Model

Provided that the multi OFDMA cellular system, bandwidth W MHz is divided into M sub-carries channel. The

paper defines M={ 1,2,…,M } as the set of physical resource blocks (PRBs). Each sub-carries bandwidth is W/M

MHz. Frequency reused factor is one. [ ]

n

m

Pis signal power of the nth cellular base station in the mth sub carrier.

1

[ ,..., ]

n n

m

P p pis each base station power vector in the mth PRB. The number of users is K. ( , )

n

k m

pdenotes power

consumed the kth user in the nth base station holds the mth PRBS. Cellular base station receives signal downlink

channel model

reduce the energy consumption by in average 80%2. Emmett Carolan et al. chose an ANN employing Back

Propagation (BP) to minimize power consumption in the cellular and the base station3. In addition, it is promising

the solution of enhancing energy efficiency to recourse management4. Energy efficient problem has been

investigated in [5-8]. Work[9] stated energy efficient allocation algorithms based on quality of server (QoS)

constrain. Although extensive studies have researched on these issues, it still lacks enough analysis and researches

on multiple antenna ultra-dense heterogeneous networks. This paper analyzes that energy efficiency, capacity and

covering problem based on 5G multiple antenna ultra-density networks. This paper is organized as follows:

section 2 introduces the architecture of networks and some attribute of networks such as multi mechanisms,

dynamic channel allocation and freely accessing. Section 3 states system model and simulation. Section 4 and 5

analyze the results and make a summary.

2．Network architecture

The 5G application environments and the network Framework are shown in fig.1 and fig.2. Obviously, the 5G

network is very complex, which is combined with the macro base station, the micro base station, the pic-cellular,

the millimeter wave (mWave) and so on. The organization form of heterogeneous and wireless self-organization

net (SON) are adopted. Therefore, this network should be flexible enough to adapt to various scenarios. It

includes 2G base station, 3G base station, 4G and LTE-A, and provides open wireless network accessing.

Fig.1. The deployment environment of the 5G7Fig.2. 5G network architecture Fig.3. Ultra density network for 5G

Fig.2 indicates 5G network architecture. The open wireless networking (OWN) combined with present cloud

technique provides more flexible for radio access. SON is a core of the whole network management. Therefore,

this architecture is called multi-mechanisms cloud accessing network (MMCLAN).

According to above description, 5G network has the following attributes:

2.1 Multi Communication Mechanisms Coexist

For satisfying the requirement of mobility, covering and capacity, solving ubiquitous and wireless accessing,

the 5G network integrates multi-communication mechanism. And it includes a variety of communication systems.

2.2 Intelligent Resource Management

Because of the complexion of network architecture, resource management confronts huge challenge. Intelligent

is a developing direction of network resource management. Increasing spectrum efficiency, saving power

consumption, improving quality of service and optimizing network need intelligent network management

technique. Therefore, 5G network provides more flexible and intelligent recourse allocation and management than

the present network.

1042 Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047

[ , ] ( , ) [ , , ] [ , , ] ( , ) [ , , ] [ , ] 0

1 1, 1

L N L

m m m m m m m

k n k n k n l k n l i j i k l i l

l i i n l

y h x h x

(2)

( , )

m

k n

adenotes the index of allocation of spectrum resource. When

( , )

m

k n

a

=1, it indicates that the kth user acquires the

mth PRB in the nth base station. Otherwise,

( , )

m

k n

a

=0.The [ , . ]

m

k n l

hdenotes the lth antenna complex channel response of

the nth BS to the kth user use the mth sub carrier. The [ , , ]

m

k n l

xis the lth antenna signal transmitted. The first item in

（2）denotes the kth user signal received. For the order of M signals (equal energy constellations),

2

| |

m m

k k

x E

,

where

m

k

E

is the energy of the signal. The second item denotes the co-channel same frequency interference from

adjacent cellular. 0

is additive white Gaussian noise (AWGN). Now, considering that

( , ) [ , ,1] [ , , ]

[ , , ]

m m m

k n k n k n L

H h h

and

( , ) [ , ,1] [ , , ]

[ , , ]

m m m H

k n k n k n L

X x x

are vectors, the received signal of kth user are

[ , ] ( , ) [ , ] [ , ] ( , ) [ , ] [ , ] 0

1,

,

N

m m m m m m m

k n k n k n k n j i k i j i

i i n

y H X H X j k

(3)

Assume that the signature sequences of the users appear as mutually uncorrelated noise. The average output

power is given by

( , ) [ , ] [ , ] [ , ] [ , ] [ , ] [ , ]

{ ( ) } { ( ) ( ) }

m m H m m m H m H

k n k n k n n k n k n k n k

e E y y E H X X H

[ , ] [ , ] [ , ] [ , ]

1,

{ ( ) ( ) }

N

m m m H m H

k i j i j i k i

i i n

E H X X H

+N0

[ , ] [ , ] [ , ] [ , ] [ , ] [ , ] 0

1,

( ) ( )

N

m m H m m H

n k n k n k i k j i i k

i i n

H H H H N

(4)

[ , ] [ , ] [ , ]

{ ( ) }

m m H

k n k n k n

E X X

is the correlation matrix of the received vector ( , )

m

k n

X, and N0 is the noise of power which the kth user

receive. The kth user’s power constraining is satisfied with

( , ) ( , ) ( , ) ,

1

( ) , ,

l

m T m m

k n k n k l k n

l

tr X X p p m n

(5)

where the n

total

pis all of user K consuming total power at the nth base station. The [ , ] [ , . ] [ , . ]

( )

m m m H

n k n k l n k l

G H His power gran of the kth

user

[ , ] [ , ]

[ , ]

0 [ , ] [ , ]

1,

( )

m

n k k n

n k N

m

i k j i

i i n

G p

p

N G p

(6)

subject to:

0

[ , ] [ , ]

( ) ,

n k n k

p n

(6-1)

The 0

[ , ]n k

denotes the lowest signal noise ratio is required. Therefore, the kth user acquired for the lowest power at

the nth base station is

0

[ , ] [ , ] 0 [ , ] [ , ] [ , ]

1,

( ) /

N

m

k n n k i k j i n k

i i n

p N G p G

(6-2)

According to (4~6) and Shannon capacity theory, the rate which the kth user gets is expressed by

Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047 1043

Chaochen Xie/ Procedia Computer Science00 (2018) 000–000

[ , ]

[ , ] [ , ]

, [ , ]

2

1

0 [ , ]

1,

log det( ( ))

log (1 )

i k

m

n k PRB n k

l l

Lk n n k

PRB N

l l

l

i k

i i n

r W I p

p G

W

N G p

(b/s) (7)

where PRB

Wis bandwidth allocated the kth user at the nth base station. The total rate of the users can be written as

[ , ] [ , ]

1 1

( , ) ,

N K

m m

n k n k

n k

R p r m

(8)

Total consumption power is given by

0

[ , ]

1 1

k

N

N

m

n k n total

n k

p P P

(9)

where P0denotes base station consumption power which includes inherently loss such as the active site cooling

and signal processing and so on. Our main objective is to maximize system rate. This is expressed by

maximize [ , ] [ , ]

1 1

( , ) ,

N K

m m

n k n k

n k

R p r m

(10)

subject to:

C1: [ , ]

0

n k

p

C2: 0

[ , ]

1

, ,

K

m n

n k n total

k

p P P m n

C3: 0

[ , ] [ , ]

( )

n k n k

p

C4: [ , ] {0,1}, , ,

m

n k n m k

C5: [ , ]

1

1,

k

N

m

n k

k

m

where C1 ,C2 ensure consume power restricted. C3 is the user signal ratio threshold. C4,C5 denote one resource

block only allocated one user at the same cellular

，

and Nkdenots the all users of the nth cellular. The objective

function (10) contains discontinuous variables such as

[ , ]

m

n k

and fractions. Therefore, the objective function (10)

is a non-linearity of the optimization problem and an non-convex optimization problem, which is hard to be

solved with traditional algorithm. We use the way of Dinkelbach to transform the original optimization into the

following optimization:

Maximize 0

[ , ] [ , ] [ , ]

1 1 1 1

( , ) * , ,

N K N K

m m n m

n k n k total n k n

n k n k

T p r P p P n m

(11)

st. C1,C3,C4,C5.In the following, we show the optimum resource allocation. From (7), we express (11) by

replacing of following:

0

[ , ] [ , ]

1 1 1

( , , )

N K N

m m

n k n k total n

n k n

T p r p P P

(12)

According to KTT condition, first let [ , ]

m

n k

=1, and make first order derivation of ( , , )H p

be zero.

[ , ] ( , ) [ , , ] [ , , ] ( , ) [ , , ] [ , ] 0

1 1, 1

L N L

m m m m m m m

k n k n k n l k n l i j i k l i l

l i i n l

y h x h x

(2)

( , )

m

k n

a

denotes the index of allocation of spectrum resource. When

( , )

m

k n

a

=1, it indicates that the kth user acquires the

mth PRB in the nth base station. Otherwise,

( , )

m

k n

a

=0.The

[ , . ]

m

k n l

h

denotes the lth antenna complex channel response of

the nth BS to the kth user use the mth sub carrier. The

[ , , ]

m

k n l

x

is the lth antenna signal transmitted. The first item in

（2）denotes the kth user signal received. For the order of M signals (equal energy constellations),

2

| |

m m

k k

x E

,

where

m

k

E

is the energy of the signal. The second item denotes the co-channel same frequency interference from

adjacent cellular.

0

is additive white Gaussian noise (AWGN). Now, considering that

( , ) [ , ,1] [ , , ]

[ , , ]

m m m

k n k n k n L

H h h

and

( , ) [ , ,1] [ , , ]

[ , , ]

m m m H

k n k n k n L

X x x

are vectors, the received signal of kth user are

[ , ] ( , ) [ , ] [ , ] ( , ) [ , ] [ , ] 0

1,

,

N

m m m m m m m

k n k n k n k n j i k i j i

i i n

y H X H X j k

(3)

Assume that the signature sequences of the users appear as mutually uncorrelated noise. The average output

power is given by

( , ) [ , ] [ , ] [ , ] [ , ] [ , ] [ , ]

{ ( ) } { ( ) ( ) }

m m H m m m H m H

k n k n k n n k n k n k n k

e E y y E H X X H

[ , ] [ , ] [ , ] [ , ]

1,

{ ( ) ( ) }

N

m m m H m H

k i j i j i k i

i i n

E H X X H

+N0

[ , ] [ , ] [ , ] [ , ] [ , ] [ , ] 0

1,

( ) ( )

N

m m H m m H

n k n k n k i k j i i k

i i n

H H H H N

(4)

[ , ] [ , ] [ , ]

{ ( ) }

m m H

k n k n k n

E X X

is the correlation matrix of the received vector

( , )

m

k n

X

, and N0 is the noise of power which the kth user

receive. The kth user’s power constraining is satisfied with

( , ) ( , ) ( , ) ,

1

( ) , ,

l

m T m m

k n k n k l k n

l

tr X X p p m n

(5)

where the

n

total

p

is all of user K consuming total power at the nth base station. The

[ , ] [ , . ] [ , .]

( )

m m m H

n k n k l n k l

G H H

is power gran of the kth

user

[ , ] [ , ]

[ , ]

0 [ , ] [ , ]

1,

( )

m

n k k n

n k N

m

i k j i

i i n

G p

p

N G p

(6)

subject to:

0

[ , ] [ , ]

( ) ,

n k n k

p n

(6-1)

The

0

[ , ]n k

denotes the lowest signal noise ratio is required. Therefore, the kth user acquired for the lowest power at

the nth base station is

0

[ , ] [ , ] 0 [ , ] [ , ] [ , ]

1,

( ) /

N

m

k n n k i k j i n k

i i n

p N G p G

(6-2)

According to (4~6) and Shannon capacity theory, the rate which the kth user gets is expressed by

1044 Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047

[ , ]

1 1

[ , ] [ , ]

[ , ]

[ , ]

[ , ] [ , ]

( , , )

N K

m

n k

n k

m

n k n k

m

m

j k

n k

m m

j n

n k n k

r

H p

p p

r

r

p p

(13)

In (13)

[ , ]

[ , ] [ , ]

[ , ]

0 [ , ]

1

ln 2

i k

m l

n k n k

PRB

N

m

l l

n k

i k

i

r G

W

pN G p

[ , ] [ , ] [ , ] [ , ]

[ , ]

0 [ , ] [ , ] 0 [ , ] [ , ]

,

1

ln 2

m n

j k j k l k j k

m

n k n n

j k j k j k j k

j l l j

r G G p

pN G p N G p

(14)

According to (14), we can get optimization solution:

[ , ] [ , ] [ , ] [ , ]

1,

[ , ]

1(1 )

ln 2

N

m l m m

n k j k j k n k

m

j j n

n n k

p G p G

t

(15)

In (15)

[ , ]

[ , ]

1,

0 [ , ] [ , ]

1,

m

Nj k

m

n k N

l

j j n

t k t k

t t j

SINR

t

N G p

(16)

The objective function of the dual problem is non-differentiable. Therefore, we use sub-gradient methods for the

Lagrange multiplier variable of non-differentiable.

0

[ , ] [ , ]

1 1

N K

m m

total n k n k n

n k

P p P

(17)

λ is the factor updating iteration process as following

0

[ , ] [ , ]

1 1

( ) ( 1)

N K

m m

total n k n k n

n k

n n P p P

(18)

According to (15), computing the power allocated of subcarrier in N base station, that is

1

{ , , }

m m

N

p p p

. Under

condition of given the power allocated, how to make the maximization capacity of signal channel of the objective

function

( , )R p

turn into the sub-problem of N×K uncorrelated to the mth subcarrier. The optimization subcarrier

allocated is:

[ , ] [ , ]

( )

( , ) arg(max ( )), ,

m

k n k n

k

m k n R p k n

(19)

( )k

is set of user at the nth cellular. The (19) denotes the mth subcarrier allocated user to maximization [ , ] [ , ]

( )

m

k n k n

R p

3.4 Optimization Algorithm

{ Begin:

Initialize: {cellular number, radius and power; users, move speed; resource number, bandwidth, lamda }

step1: divide the users into different cellular.

step2: obtain the channels state information, according to (6) ,calculate channel gain

{ while(iterNum<iterMax)

step3: under condition step1, according to (15), calculating the users’ power.

advance and retreat search algorithm and until convergence or iterNum>iterMax,

Set iterNum=iterNum+1;

Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047 1045

Chaochen Xie/ Procedia Computer Science00 (2018) 000–000

} end while;

step4: update the users power

step5: according to step 4 and (19), allocated PRBs for users.

} end;

4. Simulation analysis

This work is simulated in the Matlab (R2013a). Parameters are shown in the table 1 and table 2.

Table2:Parameters setting for simulation

Parameter Setting

Macro cellular radius 500 m

Micro cellular radius 125m

Pico-cellular radius 25m

Channel bandwidth 5 MHz

Total number of RBs 64 RBs

RB channel interval 78.125KHz

Total number of cellular 13

Users move speed 0.5m/s

Thirteen Cellular and five hundred users randomly distribute are deployed at the (1000m×1000m) area.

Different circles indicate HetNet base stations and its coverage region, and red dots are the users randomly

moving in fig.4.

-200 0 200 400 600 800 1000 1200

-200

0

200

400

600

800

1000

1200

meter/m

mter/m

cell covering and user distribution

-200 0 200 400 600 800 1000 1200

-200

0

200

400

600

800

1000

1200

meter/m

mter/m

cell covering and user distribution

0 2 4 6 8 10 12 14

0

5

10

15

20

25

30

35

40

cell power consumption

cellular

power/w

cell=13,Low SNR, PRBs=64,Bandwidth=5MHz

2 transmit antennas power consumption

4 transmit antennas power consumption

8 transmit antennas power consumption

cellular transmit power

Fig.4. The cellular and users distribution at t time Fig.5. Divided the users into different cellular Fig.6. cellular power consumption

The fig.5 shows the users divided into different cellular in according with certain strategy. The following paper

mainly discusses relation between system performance and difference antenna numbers. At the same time,

resource allocation is evaluated.Fig.6 shows each of cellular with different antenna and their consumption power.

0 2 4 6 8 10 12 14

0

1

2

3

4

5

6x 104cell system rate

cellular

rate(bit/Hz/s)

cell=13,Low SNR, PRBs=64,Bandwidth=5MHz

2 transmit antennas

4 transmit antennas

8 transmit antennas

0 2 4 6 8 10 12 14

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

cell coverage ratio

cellular

value(%)

cell=13,Low SNR, PRBs=64,Bandwidth=5MHz

2 transmit antennas

4 transmit antennas

8 transmit antennas

0 2 4 6 8 10 12 14

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

cell

energy efficency ratio(bit/s.Hz/w)

cell nergy efficency ratio

cell=13,Low SNR, PRBs=64,Bandwidth=5MHz

2 transmit antennas

4 transmit antennas

8 transmit antennas

[ , ]

1 1

[ , ] [ , ]

[ , ]

[ , ]

[ , ] [ , ]

( , , )

N K

m

n k

n k

m

n k n k

m

m

j k

n k

m m

j n

n k n k

r

H p

p p

r

r

p p

(13)

In (13)

[ , ]

[ , ] [ , ]

[ , ]

0 [ , ]

1

ln 2

i k

m l

n k n k

PRB

N

m

l l

n k

i k

i

r G

W

pN G p

[ , ] [ , ] [ , ] [ , ]

[ , ]

0 [ , ] [ , ] 0 [ , ] [ , ]

,

1

ln 2

m n

j k j k l k j k

m

n k n n

j k j k j k j k

j l l j

r G G p

pN G p N G p

(14)

According to (14), we can get optimization solution:

[ , ] [ , ] [ , ] [ , ]

1,

[ , ]

1(1 )

ln 2

N

m l m m

n k j k j k n k

m

j j n

n n k

p G p G

t

(15)

In (15)

[ , ]

[ , ]

1,

0 [ , ] [ , ]

1,

m

Nj k

m

n k N

l

j j n

t k t k

t t j

SINR

t

N G p

(16)

The objective function of the dual problem is non-differentiable. Therefore, we use sub-gradient methods for the

Lagrange multiplier variable of non-differentiable.

0

[ , ] [ , ]

1 1

N K

m m

total n k n k n

n k

P p P

(17)

λ is the factor updating iteration process as following

0

[ , ] [ , ]

1 1

( ) ( 1)

N K

m m

total n k n k n

n k

n n P p P

(18)

According to (15), computing the power allocated of subcarrier in N base station, that is

1

{ , , }

m m

N

p p p

. Under

condition of given the power allocated, how to make the maximization capacity of signal channel of the objective

function

( , )R p

turn into the sub-problem of N×K uncorrelated to the mth subcarrier. The optimization subcarrier

allocated is:

[ , ] [ , ]

( )

( , ) arg(max ( )), ,

m

k n k n

k

m k n R p k n

(19)

( )k

is set of user at the nth cellular. The (19) denotes the mth subcarrier allocated user to maximization

[ , ] [ , ]

( )

m

k n k n

R p

3.4 Optimization Algorithm

{ Begin:

Initialize: {cellular number, radius and power; users, move speed; resource number, bandwidth, lamda }

step1: divide the users into different cellular.

step2: obtain the channels state information, according to (6) ,calculate channel gain

{ while(iterNum<iterMax)

step3: under condition step1, according to (15), calculating the users’ power.

advance and retreat search algorithm and until convergence or iterNum>iterMax,

Set iterNum=iterNum+1;

1046 Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047

Chaochen Xie/ Procedia Computer Science00 (2018) 000–000

7.Gjendemsjo A, Yang H C, Alouini M S, et al. Joint Adaptive Modulation, Diversity Combining, and Power Control for Uplink

Transmission in Two-cellular Wireless Networks[C]//2007 4th International Symposium on Wireless Communication Systems. IEEE,

2007: 272-276.

8.Tombaz S, Sung K W, Zander J. On metrics and models for energy-efficient design of wireless access networks[J]. IEEE Wireless

Communications Letters, 2014, 3(6): 649-652.

9.Wang, Luhan, et al. "Open wireless network architecture in radio access network." Vehicular Technology Conference (VTC Fall), 2013

IEEE 78th. IEEE, 2013.

10.Dinkelbach W.On nonlinear fractional programming [J]. Management Science,1967, 13(7):492-498

11.Yuan, Deyu, et al. "Energy-Efficient Resource Allocation for Multi-Cellular OFDM Networks." IETE Journal of Research 61.5 (2015):

482-491.

Fig.7. Each cellular rate Fig.8. Each cellular coverage ratio Fig.9. Each cellular efficiency ratio

Simulations demonstrate that each of cellular consuming power has been raised with the cellular antenna

numbers increasing. Macro base station and micro base station consume power below their max power. Under

satisfied with user requirements condition, each of pico-cellular power consumption slightly exceeds their max

power. And these base stations consume total power below macro base’s maximization transmitting power. So

Ultra Density Network can decrease the macro base station transmitting power and save energy consumption.

Fig.7 shows the cellular throughput. When each of cellular antenna number arises from 2 to 8, simulation result

shows that the cellular rate is also arising. The fig.7 displays that relation between cellular transmitting power and

rate. To MIMO UDN system, cellular rate is highly relation with cellular power. For example, macro base station

is high transmitting power, given condition allocation users and signals noise ratio, and every user acquiring

power is high. As a result, cellular rate is higher than low power cellular, such as micro cellular and pico-cellular.

Fig.8 describes cellular user getting resource allocation real-time coverage ratio. Coverage ratio denotes

average statistics value at one moment, not theoretical value. For example, macro base station with 2 antennas

real-time coverage ration higher from 72 percent to 94 percent with 8 antennas. How to understand the

phenomenon? Authors think that massive antenna system under complication around have more advantages than

few antenna.

Fig.9 states each of cellular energy efficiency. No changing simulations condition, the results show that the

transmitting power is higher, the more energy efficiency ratio. But with cellular power decreasing, it leads to

reduce cellular energy efficiency ratio, such as micro cellular and pico-cellular. Therefore, way increasing antenna

number, cannot obtain high energy efficiency ratio in UDN low transmitting power cellular.

5. Conclusions

With the advent of 5G era, various billions of devices and users will access network. This consumes various

energy and spectrum resource. Hence, for saving energy consumption and enhancing frequency efficiency,

ultra-dense network plays as efficiency ways. Through multi antenna UDN model simulation analysis, it proves

the plan is effective in the paper. Therefore, the paper obtains that a conclusion that reasonable deployment of

cellular networks can reduce energy consumption, enhance spectrum effective and improve the comprehensive

performance of networks.

Acknowledgements

This work was partially supported was supported by the National Natural Science Foundation of China (No.

61571069) and Project No.106112017CDJQJ168817 supported by the Fundamental Research Funds for the

Central Universities.

References

1.Damiano Rapone, Dario Sabella, Maurizio Fodrini. Rapone D, Sabella D, Fodrini M. Energy efficiency solutions for the mobile network

evolution towards 5G: an operator perspective[C]//Sustainable Internet and ICT for Sustainability (SustainIT), 2015. IEEE, 2015: 1-9.

2.Dingwen Yuan, Michael Riecker, and Matthias Hollick. Yuan D, Riecker M, Hollick M. Making ‘glossy’networks sparkle: Exploiting

concurrent transmissions for energy efficient, reliable, ultra-low latency communication in wireless control networks[C]//European

Conference on Wireless Sensor Networks. Springer International Publishing, 2014: 133-149.

3.Emmett Carolan, Seamus C, McLoone and Ronan Farrell. Carolan E, McLoone S C, Farrell R. A Predictive Model for Minimising Power

Usage in Radio Access Networks[C]//International Conference on Mobile Networks and Management. Springer International Publishing,

2015: 55-67.

4.Rao J B, Fapojuwo A O. A survey of energy efficient resource management techniques for multicellular cellular networks[J]. IEEE

Communications Surveys & Tutorials, 2014, 16(1): 154-180.

5.Yang Fengyi, Wang Haining, MEI Chengli, Zhang Jianmin, Wang Min. Yang F, Wang H, Mei C, et al. A flexible three clouds 5G mobile

network architecture based on NFV & SDN[J]. China Communications, 2015, 12(Supplement): 121-131.

6.Coskun C C, Davaslioglu K, Ayanoglu E. An Energy-Efficient Resource Allocation Algorithm with QoS Constraints for Heterogeneous

Networks[C]//2015 IEEE Global Communications Conference (GLOBECOM). IEEE, 2015: 1-7.

Xie Chaochen et al. / Procedia Computer Science 131 (2018) 1039–1047 1047

Chaochen Xie/ Procedia Computer Science00 (2018) 000–000

7.Gjendemsjo A, Yang H C, Alouini M S, et al. Joint Adaptive Modulation, Diversity Combining, and Power Control for Uplink

Transmission in Two-cellular Wireless Networks[C]//2007 4th International Symposium on Wireless Communication Systems. IEEE,

2007: 272-276.

8.Tombaz S, Sung K W, Zander J. On metrics and models for energy-efficient design of wireless access networks[J]. IEEE Wireless

Communications Letters, 2014, 3(6): 649-652.

9.Wang, Luhan, et al. "Open wireless network architecture in radio access network." Vehicular Technology Conference (VTC Fall), 2013

IEEE 78th. IEEE, 2013.

10.Dinkelbach W.On nonlinear fractional programming [J]. Management Science,1967, 13(7):492-498

11.Yuan, Deyu, et al. "Energy-Efficient Resource Allocation for Multi-Cellular OFDM Networks." IETE Journal of Research 61.5 (2015):

482-491.

Fig.7. Each cellular rate Fig.8. Each cellular coverage ratio Fig.9. Each cellular efficiency ratio

Simulations demonstrate that each of cellular consuming power has been raised with the cellular antenna

numbers increasing. Macro base station and micro base station consume power below their max power. Under

satisfied with user requirements condition, each of pico-cellular power consumption slightly exceeds their max

power. And these base stations consume total power below macro base’s maximization transmitting power. So

Ultra Density Network can decrease the macro base station transmitting power and save energy consumption.

Fig.7 shows the cellular throughput. When each of cellular antenna number arises from 2 to 8, simulation result

shows that the cellular rate is also arising. The fig.7 displays that relation between cellular transmitting power and

rate. To MIMO UDN system, cellular rate is highly relation with cellular power. For example, macro base station

is high transmitting power, given condition allocation users and signals noise ratio, and every user acquiring

power is high. As a result, cellular rate is higher than low power cellular, such as micro cellular and pico-cellular.

Fig.8 describes cellular user getting resource allocation real-time coverage ratio. Coverage ratio denotes

average statistics value at one moment, not theoretical value. For example, macro base station with 2 antennas

real-time coverage ration higher from 72 percent to 94 percent with 8 antennas. How to understand the

phenomenon? Authors think that massive antenna system under complication around have more advantages than

few antenna.

Fig.9 states each of cellular energy efficiency. No changing simulations condition, the results show that the

transmitting power is higher, the more energy efficiency ratio. But with cellular power decreasing, it leads to

reduce cellular energy efficiency ratio, such as micro cellular and pico-cellular. Therefore, way increasing antenna

number, cannot obtain high energy efficiency ratio in UDN low transmitting power cellular.

5. Conclusions

With the advent of 5G era, various billions of devices and users will access network. This consumes various

energy and spectrum resource. Hence, for saving energy consumption and enhancing frequency efficiency,

ultra-dense network plays as efficiency ways. Through multi antenna UDN model simulation analysis, it proves

the plan is effective in the paper. Therefore, the paper obtains that a conclusion that reasonable deployment of

cellular networks can reduce energy consumption, enhance spectrum effective and improve the comprehensive

performance of networks.

Acknowledgements

This work was partially supported was supported by the National Natural Science Foundation of China (No.

61571069) and Project No.106112017CDJQJ168817 supported by the Fundamental Research Funds for the

Central Universities.

References

1.Damiano Rapone, Dario Sabella, Maurizio Fodrini. Rapone D, Sabella D, Fodrini M. Energy efficiency solutions for the mobile network

evolution towards 5G: an operator perspective[C]//Sustainable Internet and ICT for Sustainability (SustainIT), 2015. IEEE, 2015: 1-9.

2.Dingwen Yuan, Michael Riecker, and Matthias Hollick. Yuan D, Riecker M, Hollick M. Making ‘glossy’networks sparkle: Exploiting

concurrent transmissions for energy efficient, reliable, ultra-low latency communication in wireless control networks[C]//European

Conference on Wireless Sensor Networks. Springer International Publishing, 2014: 133-149.

3.Emmett Carolan, Seamus C, McLoone and Ronan Farrell. Carolan E, McLoone S C, Farrell R. A Predictive Model for Minimising Power

Usage in Radio Access Networks[C]//International Conference on Mobile Networks and Management. Springer International Publishing,

2015: 55-67.

4.Rao J B, Fapojuwo A O. A survey of energy efficient resource management techniques for multicellular cellular networks[J]. IEEE

Communications Surveys & Tutorials, 2014, 16(1): 154-180.

5.Yang Fengyi, Wang Haining, MEI Chengli, Zhang Jianmin, Wang Min. Yang F, Wang H, Mei C, et al. A flexible three clouds 5G mobile

network architecture based on NFV & SDN[J]. China Communications, 2015, 12(Supplement): 121-131.

6.Coskun C C, Davaslioglu K, Ayanoglu E. An Energy-Efficient Resource Allocation Algorithm with QoS Constraints for Heterogeneous

Networks[C]//2015 IEEE Global Communications Conference (GLOBECOM). IEEE, 2015: 1-7.