Download A New Model of Genetic Zone Routing Protocol

A New Model of Genetic Zone Routing Protocol (GZRP) :
The Process of Load Balancing and Offloading on The
UMTS - IEEE 802.11g Hybrid Networks
Setiyo Budiyanto
Department of Electrical Engineering, Universitas Mercu Buana, Jakarta, Indonesia
Email : [email protected]
Abstract - The stages of the process of Genetic
Algorithm (GA), are: Encoding Genotype and
Chromosome; Set Initialization Population; Evaluation
Fitness Function; and Selection Process as well as in
the later stages Cross Over Process and Mutation.
Outputs from the tests performed in this study can be
obtained by comparing the Genes of the "Child"
(condition data traffic on the UMTS Hybrid - 802.11g
network after the GA) against Gen "Holding" (traffic
data before the GA process).
The research was conducted by calculating the
environmental factors, namely: The scheme Two - Ray
Model Propagation and Overlapping Channel
Interference Factor, the Doppler Effect be ignored
because the User Equipment (UE) is considered not to
shift significant arenas on the IEEE 802.11g networks.
The results of the research is as follows: In the process
of cross over, there is a significant change in the
bandwidth, data traffic capacity and Power parameter
changes by 9 MHz, 36 MB, and 40 dBm. In the process
of mutation, there is a significant change in the
bandwidth, data traffic capacity, and Power
parameter by 17 MHz, 32 MB, and 20 dBm.
Keywords : Genetic Algorithm, UMTS - IEEE 802.11g
Hybrid Networks, Two - Ray Model Propagation,
Overlapping Channel Interference Factor
2
Introduction
The utilization of data traffic on the Universal Mobile
Telecommunications System (UMTS) has been increased,
along with the supporting technology development [1]. It
resulted in the idea to stream the mobile data traffic to
other wireless networks such as Wireless Fidelity, in this
study used tissue type IEEE802.11g[2-3]. Mobile
Advanced Delivery Network (MADNETs) is the network
is able to process the transfer of data from several different
networks (UMTS and WIFI)[4-5]. The process of moving
data traffic from UMTS network to a WiFi network and
also the reverse process (the transfer of data traffic from
the WiFi network to a UMTS network) is namely the process
of UMTS - WiFi Offload. Vertical Hand Over (VHO)
algorithm Used in order to perform the offloading process [67]. Until nowadays, there has not been a study that addresses
the special routing protocols that can divert traffic to other
access points that are in a cluster. Meanwhile, on the other
hand technology Mobile Ad-Hoc Network (MANET) has
some concept of routing protocols : Reactive, Proactive, and
Hybrid [8].
The activities carried out at the investigation are:
Development of Genetic Zone Routing Protocol (GZRP) that
has been developed in previous studies [9-11]. Namely by
elaborating the Genetic Algorithm begin from population
initialization process with the best individuals search by cross
over methods as well as to the process of genetic mutation.
1
Related Research
The development of technology resulted in increased usage of
data in the mobile network. On the other hand, the technology
that existed at present incapable of becoming the solution of
the problem of increased traffic on these data, it brought the
idea to drain the mobile traffic to other wireless networks such
as WiFi [12-13]. A MADNET technology architecture that is
capable of forwarding traffic to mobile wireless networks
such as WiFi accordance research conducted in [4]. On the
process of offloading traffic from UMTS network to a WiFi
network, contained VHO algorithm is : capable of performing
the handover process from the UMTS network to a WiFi
network [6]. In UMTS - WiFi Offload technology there has
been no special routing protocols that can divert traffic to
other access points that are in a cluster. On the other hand,
wireless technologies such as Mobile Ad-Hoc Network
(MANET) has some concept of routing protocols : Reactive
(on-demand), Proactive, and Hybrid [8].
In [10-13], a concept known as load balancing using Genetic
Zone Routing Protocol (GZRP) on the wireless network
MANET, which is a load balancing with bringing the total
packets received to an alternative route so as to reduce the
traffic load on a standalone service. According to [9], GZRP
is the development of hybrid routing protocols in MANET
1
namely Zone Routing Protocol (ZRP) which is coupled with
Genetic Algorithm (GA) [14]. ZRP is used to reduce the
burden of proactive routing protocol control and reduce the
latency caused by the discovery in a reactive routing
protocol. In [15] described a WiFi network interworking
and UMTS networks, one of which is the handover process
that goes from the cellular network to the WLAN. In [6]
described a concept handover algorithms set up with both
the process of handover from UMTS to WiFi known as
algorithms Vertical Handover (VHO). In [6], Vertical
Handover (VHO) focused on Goodput and RSSI, known as
Hybrid-RSSI algorithm. One thing that is not easy to get an
accurate estimate of goodput in the real environment. The
second proposal is a hybrid-RSSI estimation algorithms
utilizing VHO goodput to ensure the quality of services
provided to users.
3
Proposed Algorithm
Offloading process occurred conditions at the time a UMTS
RSSI value less than the RSS threshold (carried out process
of moving the traffic from a UMTS networks to WiFi). The
parameters used in this study were related to load balancing
using GZRP (in the WiFi network) is the bandwidth,
capacity and power. These three parameters will be linked
with other methods of dealing with the propagation and
interference, while the effect doopler ignored for User
Equipment (UE) is considered not to shift a significant
place. Propagation models used is the: Two-ray propagation
model models, while the interference parameters using
Overlapping Channel Interference Factor. The algorithms
developed in the study can be seen in Figure 1.:
In [10-11] [16-17], has carried out research related to the use
of protocols GZRP combined with VHO occupancy in order
to achieve efficiency levels of data traffic, either on the
network or UMTS and WiFi networks; as well as the
simultaneous performance of both networks. The discussion
in detail to the discussion of performance improvement
GZRP protocol can be used as a very interesting discussion
in order to seek more comprehensive solutions related to
problems that occur in the process of load balancing using
GZRP protocol..
The Hybrid protocol used is the Genetic Zone Routing
Protocol (GZRP) [10-11], is able to improve the efficiency
of the performance of a network. It can also reduce the load
on the track by balancing the distribution of packet delivery
over the course of the alternatives available. Utilization
concept GZRP especially in the process of traffic load
balancing between nodes.
This study aims to investigate intelligent systems load
balancing algorithms and WiFi offload in the wireless
communication traffic bottlenecks can be overcome; in
addition, also equal distribution of traffic and to develop
methods of genetic algorithm combined routing protocol
ZRP (Zone Routing Protocol) in the process of offloading in
the load balancing and IEEE 802.11g WiFi network. In this
case the load balancing process in the Hybrid UMTS
networks - IEEE 802.11g. Load balancing happened in the
WiFi network is a follow up of the process of offloading
data traffic from UMTS network to a WiFi network.
Offloading process occurs when the condition value is
smaller than the UMTS RSS RSS threshold (do the process
of moving traffic from UMTS networks to WiFi).
START
Inactive
no
no
RSS >
Threshold
no
RSS >
Threshold
yes
yes
WiFi Active
UMTS Active
yes
RSS >
Threshold
yes
Link Going
Down Event
Mx = 0
yes
yes
Mx = 0
no
no
RSS >
Threshold
no
Data Transmit
no
no
no
Data Transmit
Capacity Fall
Or Status Down
yes
T = 50s
GZRP
T = 50s
yes
END
yes
Figure.1. The Proposed Algorithm
Figure 1. Schematic of a working system algorithm
developed in the the research, it can be seen Wi-Fi offload
process that can be described as follows:
 User Equipment (UE) is automatically connected to
the UMTS network, if RSS> RSS threshold, then it
would use the UMTS network; If the RSS <RSS
threshold then alert will be given a link going down
and will carried out Initialization network
2


conditions and checking RSS in the nearest WiFi
node.
If the WiFi network also did not have RSS> RSS
threshold, then the demand for packet was
suspended for a specified timeout. Subsequently,
the EU network be deactivated.
At the time of the network used, has one of the
parameter value 0 of the variable Mx, it will be
vertical handover. The parameters that determine
handover is defined in equation (1).
𝑀𝑥 = 𝑓(𝑏𝑥 − 𝑏𝑡ℎ ) . 𝑓(𝑅𝑆𝑆𝑥 − 𝑅𝑆𝑆𝑡ℎ ) . (𝑃𝑥 −
𝑃𝑡ℎ ) . 𝑓(𝐶𝑥 − 𝐶𝑡ℎ )
Annotations :
Pt
=
Power Transmitter
r1
=
Direct distance from Tx to Rx
r2
=
Non-direct distance from Tx to Rx
Γ(α)
=
The reflection coefficient depends on the
angle α and polarization
Where as,
(1)
K(two ray) = Gt Gr ht2 hr2 / L
(3)
Annotations :
𝑏𝑥
= The bandwidth used by the network to connect
(to a UMTS network or WiFi network)
bth
= The threshold of the bandwidth required by the
EU
= The value of the received signal strength by the
EU
= The RSS threshold value of the EU
𝑅𝑆𝑆
𝑅𝑆𝑆𝑡ℎ
𝑃𝑥
Pth
𝐶𝑥
Cth
= The energy possessed by the base station and
the access point
= The energy required by the EU
= The capacity of the network service capabilities
by the EU
= The threshold of capacity is required to process
UE connection to the network
In a WiFi network, are applied the algorithm with the ability
to determine when capacity nodes is in the process of
transition or receive the data reaches the threshold or link
node status is "down" applied GZRP routing protocol.
The parameters of the test conducted
The novelty of the system testing parameters Model Genetic
Zone Routing Protocol (GZRP) in the IEEE 802.11g
networks; It is the process of load balancing and offloading
in the Hybrid a UMTS - 802.11g networks composed by
some of the main discussion, namely: Two-Ray models
Propagation, Overlapping Channel Interference Factor and
processes on the network GZRP MADNETs. The following
discussion related of testing parameters:
Two-Ray Model Propagation
Propagation be one of the testing parameters in order to
consider factors existing environment. This parameter is
used in the testing by considering previous studies [18-22].
In this test used models Two-Ray Propagation Model. The
Power receive signals (Pr) on the propagation of two - ray
models, defined in equation (2) :
𝜆
1
1
4𝜋
𝑟1
𝑟2
𝑃𝑟 = 𝑃𝑡 ( ) [ 𝑒 (𝑗𝑘𝑟1) + 𝛤(𝛼)
𝑒 (−𝑗𝑘𝑟2) ]
2
(2)
Gt and Gr
=
Gain Tx and Rx
ht and hr
=
Antenna high Tx and Rx
λ
=
Wave Length (λ= c/f)
L >=1, loss factor at the systems (L=1)
𝛤(𝜃) =
cos 𝜃− 𝛼√𝜖𝑟−𝑠𝑖𝑛2 𝜃
cos 𝜃+ 𝛼√𝜖𝑟−𝑠𝑖𝑛2 𝜃
(4)
Annotations :
θ
=
90-α
a
=
1/ε or 1 for polaritation vertical or horizontal
r
=
The relative dielectric constant in the reflective
surface
Overlapping Channel Interference Factor
Model overlapping channel interference is to determine
factor overlapping channel interference; wij be the
percentage increase in value relative in the interference
generated by two access points i and j using overlapping
channels [23-25]; The next channel is given to the AP must
be chosen a thorough. Method aims overlapping channels to
minimize channel interference from several different AP, it
is the development of research assumes that some APs
occupy the same channel. As we know that interference in
the communication network greatly affects the value of in
the AP power transmitter and power receiver in the UE.
The value of the overlapping channel interference factor can
be determined by taking into account the equation (5) as
follows:
1 − |𝐹𝑖 − 𝐹𝑗 |𝑥 𝐶 𝑤𝑖𝑗 ≥ 0
𝑤𝑖𝑗 = {
0
𝑜𝑡ℎ𝑒𝑟𝑠
Annotations :
𝐹𝑖
𝐹𝑗
(5)
= channel assigned to the access point i
= channel assigned to the access point j
3
𝐶
= Overlapping channel factor
4
Simulation, Results and Discussion
The study bounded during load balancing from a UMTS
network to the IEEE 802.11g network. The Genetic
Algorithm process begins; As for the stages of the process is
carried out as follows: Encoding, Population Initialization,
Determination of Value Fitness, Selection, Cross Over and
Mutation.
WiFi
To initiate this process first define the problem, the scenario
was made of the condition of User Equipment (UE) and
Access Point (AP). AP assumed each of which there are 10
with the bandwidth, capacity and power as follows:
UE
AP 2
AP 3
AP 4
AP 5
AP 6
13
25
20
AP7
14
30
20
AP8
16
40
20
AP9
18
45
20
AP10
20
54
20
Details of the process, described further, as follows:
Encoding in the process of Genetic Algorithm
The encoding process in the genetic algorithm is the process
to encode a gene in a chromosome. These genes can be
represented in the form of binary bits, real number, a list of
rules, permutation elements, elements of the program or
other representation that can be operated in the genetic
operations.
In load balancing, to a WiFi network each node is encoded
form of binary numbers are adjusted for the number of
nodes. Forms for each individual chromosome can be seen
in Table 2, as follows:
Power 10 dBM
Video via Youtube 5MB
-GZRP-
AP 1
AP6
Table 2. The form of chromosomes to each individual (in
binary format)
AP 7
AP 8
AP 9
AP 10
Figure 2. Assumptions for User Equipment (UE)
Annotations :
Bandwidth (Bth) = 5 MHz
Capacity (Cth)
= 5 MB
Power (Eth)
= 10 dBm
The value of the parameters contained in each access point
can be seen in Table 1, as follows:
Table 1. Bandwidth, Capacity and Power in the AP
Bandwidth
(MHz)
Capacity
(MB)
Power
(dBm)
AP1
5
6
20
AP2
6
8
20
AP3
8
10
20
AP4
9
15
20
AP5
11
20
20
Bandwidth
(MHz)
Capacity
(MB)
Power
(dBm)
AP1
000101
000110
010100
AP2
000110
001000
010100
AP 3
001000
001010
010100
AP4
001001
001111
010100
AP5
001011
010100
010100
AP6
001101
011001
010100
AP7
001110
011110
010100
AP8
010000
101000
010100
AP9
010010
101101
010100
AP10
010100
111000
010100
Based on the data presented in Table 1, it can be seen that
the WiFi network (IEEE 802.11g) consisted of 10
individuals (access point) where each chromosome is
defined in binary form. Defining chromosomes are based on
three parameters such as bandwidth that put the group first
column, column group capacity in the second and third
4
power in the column. Each chromosome is made up of six
genes that total in a single individual genes to 18 genes.
Based on the simulation parameters used the number of
nodes that there are 10 nodes AP as shown in pseudo code as
which can be seen in the section below:
𝐸𝑥𝑖
=
Power of Access Point
𝐸𝑡ℎ
=
Power threshold (contained in the UE)
The following pseudocode for the formation a fitness
function that has been presented at the equation (6):
void fitness ()
{char *fitness = malloc (sizeof (*fitness)
*Num_Elements)
# AddParams set cluster_num_
# AddParams set nodes_num_
for (i=0; i <nn; i++)
set node_(0) [$ns node]
{ if(bandwidth_remain > bandwicdthTh)
set node_(1) [$ns node]
{A=1}
set node_(2) [$ns node]
else {A=0}
set node_(3) [$ns node]
if(Capacity_remain > CapacityTh)
set node_(4) [$ns node]
{B=1}
set node_(5) [$ns node]
else {B=0}
set node_(6) [$ns node]
if(RSSI_remain > RSSIth)
set node_(7) [$ns node]
{C=2}
set node_(8) [$ns node]
else {C=0}
set node_(9) [$ns node]
X[i] = [A*(Bxi – Bth) + B*(Cxi – Cth) + 1
/ (C*(Exi – Eth))]
Fitness Function in The Genetic Algorithm
fitness = sum (X[i])
A function is a reference for the optimal value based on the
objectives, namely: the selection of the optimal AP for the
EU to consider the bandwidth, capacity and power. The
fitness function used can be seen in equation (6) as follows:
}
𝑋[𝑖] = [𝐴 ∗ (𝐵𝑥𝑖 − 𝐵𝑡ℎ ) + 𝐵 ∗ (𝐶𝑥𝑖 − 𝐶𝑡ℎ ) + (
1
𝐶∗(𝐸𝑥𝑖 −𝐸𝑡ℎ
)]
(6)
Description of the parameters used in equation (6):
X[i]
=
A
=
Fitness Function for the access point number i
Coefficient of Bandwidth
B
=
Coefficient of Capacity
C
=
Coefficient of Power Consumption
𝐵𝑥𝑖
=
Bandwidth contained in the access point
𝐵𝑡ℎ
=
Bandwidth Threshold (contained in the UE)
𝐶𝑥𝑖
=
Capacity of Access Point
𝐶𝑡ℎ
=
Capacity Threshold (contained in the UE)
Results and Analysis of The Research
In this section will be discussed related to the results and
analysis of outcomes which have been obtained. In this
research, discusses the process of genetic algorithms starting
from the stage of the gene encoding the individual
constituent (access point) and ending with the mutation of an
existing gene. The following discussion details about the
results of research and analysis of the results of this study:
Process of Cross Over at the MADNETs Network by
GZRP
The discussion done with the assistance of software NS2.35
and Matlab; for further validation of the data based on by
parameter adjustment (6). Be discussed in more details
regarding the parameters of bandwidth, capacity and power
before and after through GZRP (by comparing the value of
both of them using the parameter propagation and
interference).
Parameter Bandwidth in the Crossover Process
The first parameter is used as a media data validation at the
the cross over is bandwidth. Values bandwidth before
experiencing GZRP process is the result of the bandwidth
5
that has been presented in Table 1 while the value of the
bandwidth after processing GZRP is the value of the match
by fitness on the bandwidth which has been interbred by
another access point.
Outcomes bandwidth obtained before and after the process
of GZRP, as shown in Figure 3:
Based on the figure 4., can be seen the results of the cross
over process in the capacity of traffic data parameters
contained in the access point. It was concluded that the
access point 5 has a value change is the greatest capacity of
traffic data is 36 MB.
Parameter Power in the Crossover Process
The value of power before experienced the process GZRP is
the result of a power that has been presented in Table 1.,
while the value of power after processing GZRP is the value
of the match with fitness on the part of the capacity that has
undergone cross over to the other access point.
Outcomes of power earned before and after the process of
GZRP, as shown in Figure 5:
Figure 3. Bandwidth on the conditions before and after the
process GZRP
Based on the figure 3., can be seen the results of the cross
over process in the bandwidth parameters contained in the
access point. It was concluded that the access point 5 has a
value change is the greatest bandwidth is 9 MHz.
Parameter Capacity Data Traffic in the Crossover
Process
In the cross over process, the data traffic capacity is used as
a parameter in terms of measurements and data validation.
The value of capacity in before the GZRP process is the
result of the capacity that has been presented in table 1,
while the value of capacity after processing GZRP is the
value of the match with fitness on the part of the capacity
that has undergone cross over to the other access point.
Outcomes of capacity earned before and after the process of
GZRP, as shown in Figure 4.
Figure 4. Capacity Data Traffic on the conditions before and
after the process GZRP
Figure 5. Power Bandwidth on the conditions before and
after the process GZRP using "two ray model" and
interference "dynamic overlapping channels"
Based on the figure 5., can be seen the results of the cross
over process in the power parameters contained in the access
point. It was concluded that the access point 5 has a value
change is the greatest power is 40 dBm.
Process of Mutation at the MADNETs Network by
GZRP
Mutations in the the GA will influence the gene at the an
individual, which generates the best individual or otherwise
decreased of quality. In the process of testing mutations in
MADNETs network using GZRP, which used the same
parameters as in the process of Cross Over : Bandwidth,
Capacity Data Traffic and Power.
Parameter Bandwidth in the Mutation Process
The first parameter is used as a tool for validation data on
the mutation process is bandwidth. The value of bandwidth
before experiencing the GZRP process is the result of the
bandwidth that has been presented in table 1.; The value of
Bandwidth after processing GZRP is the value of the
matches with fitness on the bandwidth that has develop of
mutation.
Outcomes of bandwidth earned before and after the process
GZRP, can be seen in Figure 6:
6
Parameter Power in the Mutation Process
The value of power before experienced the process GZRP is
the result of a power that has been presented in Table 1.,
while the value of power after processing GZRP is the value
of the match with fitness on the part of the capacity that has
undergone mutation to the other access point.
Outcomes of power earned before and after the process of
GZRP, as shown in Figure 8:
Figure 6. Bandwidth on the conditions before and after the
process GZRP
Based on the figure 6., can be seen the results of the cross
over process in the bandwidth parameters contained in the
access point. It was concluded that the access point 8 has a
value change is the greatest bandwidth is 17 MHz.
Parameter Capacity Data Traffic in the Mutation
Process
In the mutation process, the data traffic capacity is used as a
parameter in terms of measurements and data validation. The
value of capacity in before the GZRP process is the result of
the capacity that has been presented in table 1, while the
value of capacity after processing GZRP is the value of the
match with fitness on the part of the capacity that has
undergone mutation to the other access point.
Outcomes of capacity earned before and after the process of
GZRP, as shown in Figure 7.
Figure 8. Power Bandwidth on the conditions before and
after the process GZRP using "two ray model" and
interference "dynamic overlapping channels"
Based on the figure 8., can be seen the results of the cross
over process in the power parameters contained in the access
point. It was concluded that the access point 5 has a value
change is the greatest power is 20 dBm.
5
Conclusion
In this section, there will be a discussion and analysis of
results obtained from testing the model renewal GZRP on
IEEE 802.11g network, the load balancing the process and
offloading on Hybrid UMTS networks - IEEE 802.11g. The
process of discussion and analysis of the test results of the
research that has been done is to determine the performance
of the performance of the system that has been created based
on the methodology proposed in this study.
1.
2.
Figure 7. Capacity Data Traffic on the conditions before and
after the process GZRP
Based on the figure 7., can be seen the results of the
mutation process with a focus on parameter capacity data
traffic at the the access point. It was concluded that the
access point 5 has increased value changes most capacity is
equal to 20 MB, while AP 8 has a change of impairment
greatest capacity that is equal to 32 MB.
3.
The process of GZRP on the MADNETs network;
The steps undertaken in this process starting from
the
Encoding,
Population
Initialization,
Determination of Value Fitness, Selection, Cross
Over and mutation.
The discussion on the results of research that has
been done by calculating the propagation factor
"two ray model" and interference "dynamic channel
assignment".
In the the process crossovers; Access point 5 have
the value changes bandwidth, capacity and power
of the greatest in the amount of 9 MHz, 36 MB and
40 dBm. Retrieved a conclusion based on these
conditions, namely: User Equipment will be
connected to the Access Point 5 at the time of the
7
4.
selection of the access point with the network
UMTS - WiFi (IEEE 802.11g).
In the process of mutation; Access point 5 have the
value changes bandwidth, capacity and power of
the greatest in the amount of 17 MHz, 20 MB and
20 dBm MHz. Retrieved a conclusion based on
these conditions, namely: User Equipment will be
connected to the Access Point 5 at the time of the
selection of the access point with the network
UMTS - WiFi (IEEE 802.11g).
[9]
P. Sateesh Kumar and S. Ramachandram, “Genetic
Zone Routing Protocol”, Int. Jrl. of Theoretical and
appl. Infn. Tech., 4(9), pp 789- 794, Sept. 2008
[10]
Budiyanto, S., Asvial, M., danGunawan, D.,”
Implementation of Genetic Zone Routing Protocol
(GZRP) in UMTS-WiFi Offload”, Institute of
Electical and Electonics Engineers (IEEE) Tencon
2013”, pp. 72 – 77, Xi’an – China 22 – 25 Oktober
2013
[11]
Budiyanto, S., Asvial, M., dan Gunawan, D.,”
Performance Analysis of Genetic Zone Routing
Protocol Combined with Vertical Handover
Algorithm in UMTS – WiFi Offload”, Journal of
ICT Research and Applications Vol. 8 no.1 pp. 49 –
63, Mei 2014
6
References
[1]
Cisco , “Cisco Visual Networking Index: Global
Mobile Data Trafik Forecast Update, 2012–2017”,
White Paper, USA, Feb. 2013
[2]
B. Han, P. Hui, V. S. A. Kumar, M. V. Marathe, G.
Pei, and A. Srinivasan,”Cellular Traffic Offloading
through Opportunistic Communications: A Case
Study”, Proceedings of the 5th ACM workshop on
Challenged networks, pages 31 – 38, 2010
[12]
Hassan al-mahdi et.al. ”Performance Evaluation of
some Routing Protocols using TCP Traffic Types in
Mobile Ad Hoc Networks” internasional journal of
computer science issues, vol. 11, Issues 4, No.
Minia university. Eqypt. July 2014
[3]
J. Lee, Y. Yi, S. Chong, and Y. Jin,”Economics of
WiFi
Offloading:
Trading
Delay
for
CellularCapacity”, Proceeding of Canadian Organ
Replacement Register (CORR), pages 1 – 15, Dec
31, 2012
[13]
P. S. Kumar, and S. Ramachandram, “Load
Balancing in Genetic Zone Routing Protocol for
MANETs”, International Journal of Computer and
Information Engineering Vol. 3:4, pages 261 – 266,
2009
[4]
Chandrasekhar, V., Andrews, J.G., & Gatherer, A.,
“Femtocell
networks:
A
survey,”
IEEE
Communications Magazine, vol. 46, no. 9, pp. 59–
67, Sept. 2008
[14]
Goldberg. D. E, “Genetic Algorithms in Search,
Optimization and Machine Learning”, AddisonWesley, New York, NY 1989
[5]
Dimatteo, S., Hui, P., Han, B., &Li, V. O. K., ”
Cellular Trafik Offloading through WiFi
Networks”, 8th IEEE International Conference on
Mobile Adhoc and Sensor Sistems (MASS), pages
192 – 201, Oct 17 – 22, 2011
[15]
R. S. Kumar, and S. Ramachandram, “Scalability of
Network Size on Genetic Zone Routing Protocol
for MANETs”, International Journal of Computer
and Information Engineering, 2008
[16]
Asvial, M., Budiyanto, S., dan Gunawan, D., “An
Intelligent Load Balancing and Offloading in 3G –
WiFi Offload Network Using Hybrid and Distance
Vector Algorithm”, IEEE Symposium on Wireless
Technology and Applications (ISWTA 2014)”,pp.
36 – 40
[17]
Budiyanto, S., Asvial, M., dan Gunawan, D.,
“Implementation Dedicated Sensing Receiver
(DSR) in 3G - WiFi Offload”, International
Conference on Smart Green Technology in
Electrical and Information System (ICSGTEIS
2014),pp. 37 – 42
[18]
Ikegami,F.,
Yoshida,S.,
Takeuchi,T.
and
Umehira,M., “Propagation Factors Controlling
Mean Field Strength on Urban Streets”, IEEE
Trans. on Antennas and Propagation, Vol. AP-32,
No. 8, Aug. 1984, pp. 882-889
[19]
Rao, D. N., Murthy, M. J .K., Sarkar, S. K.,
Pasricha, P. K., Dutta, H. N. and Reddy, B. M.,
[6]
Busanelli, S., Martalo, M., Ferrari, G., &Spigoni,
G., “Vertical Handover between WiFi and UMTS
Networks: Experimental Performance Analysis”.
Internatioanl Journal of Energy, Informationand
Communication. Vol. 2, Issue 1, Feb. 2011
[7]
Johann Marquez-Barja, Carlos, T.,Calafate, Juan
Calos Cano & Pietro Manzoni, “An Overview of
Vertical Handover Techniques: Algorithms,
protocols and tools”. Computer Communication, 34
(2011) 985-987
[8]
Savita Gandhi, Nirbhay Chaubey, Naren Tada,
Srushti Trivedi, “Scenario-based Performance
Comparison of Reactive, Proactive & Hybrid
Protocols in MANET,” International Conference on
Computer Communication and Informatics (ICCCI
-2012), Jan 10 – 12, 2012
8
“Hilly Terrain LOS Fadeouts and Fresnel Zone
Considerations from Ray Tracing Techniques”,
IEEE Trans. on Antennas and Propagation, Vol.
AP-35, No. 11, Nov. 1987, pp. 1330-1333.
[20]
Anderson.
I,
Rappaport.T,
Yoshida,
S.
“Propagation Measurements and models for
wireless
communications
channels”.
IEEE
Communications magazine, januari 1995
[21]
Tapan K. Sarkar, et al., A Survey of Various
Propagation Models for Mobile Communication,
IEEE Antennas and Propagation Magazine, Vol.
45, No. 3, June 2003
[22]
M.J. Gans, “ A power-spactral theory of
propagation in the mobile radio environment”,
IEEE
Transactions
on
vehicular
technology, 21 (1), 27-38, 1972
[23]
Robert Akl, Anurag Arepally, “Dynamic channel
Assignment in IEEE 802.11 networks”, IEEE 14244-1039-8, 2007
[24]
A. Hills and R. Friday, “Radio resource
management in wireless LANs,” IEEE Radio
Communications, pp. S9–S14, December 2004
[25]
Sendra Sandra, Garcia Miguel, Turlo Carlos, Lloret
Jaime. “WLAN IEEE 802.11a/b/g/n Indoor
Coverage and Interference Performance Study.
International Journal on Advances in Networks and
Services, Volume 4 No 1 & 2, 209 – 222, 2011
9