Content uploaded by Khulood Omran
Author content
All content in this area was uploaded by Khulood Omran on Feb 04, 2021
Content may be subject to copyright.
Proc. of the 2nd International Conference on Electrical, Communication and Computer Engineering (ICECCE)
14-15 April 2020, Istanbul, Turkey
978-1-7281-7116-6/20/$31.00 ©2020 IEEE
The Routing Control in Mobile Ad hoc Network
Using Intelligent Optimization Algorithms
Khulood Moosa Omran
Electrical Engineering department
University of Basrah
Basrah, Iraq
khuloodmo@yahoo.com
Abstract—MANET (Mobile Ad hoc Network) made up of a
combination of nodes that are randomly moving with a
certain speed in any direction without relying on the
communication infrastructure. It is useful for creating a low-
cost network and communication during a disaster and to
avoid the failure of communication. It is necessary to
improve many of the criteria metrics that affect
performance. In the traditional the protocol of routing
AODV (Ad hoc On-Demand Distance Vector), the data
packets are sent toward the adjacent node only by the
shortest path method and cannot meet the multi-objective
approach. This article utilized the FOA (Fruit fly Optimized
Algorithm) to control the problem and find a suitable
routing instead of the shortest method. The results found is
compared with the particle swarm optimization (PSO)
approach and traditional AODV routing protocol, the
proposed (FOA) method offers the fastest and most accurate
pathway. The numerical simulations indicate that the
suggested approach achieved better performances in the case
of delay time and improve the efficiency of the system.
Keywords— AODV, MANET, PSO, FOA
I. INTRODUCTION
MANET is one of the wireless communications fields,
which consists of a set of mobile nodes or a collection of
communicated devices without using a specified and
central infrastructure [1]. This type of network has many
research interests for use in the military field and in
emergency and natural disasters, these networks need to
implement routing protocols that ensure that messages
reach the desired destination and achieve the goal of the
application. In this paper, it was expanded a dissection of
the routing performance for a number of the most
important routing protocols used in these networks,
namely AODV, Quality of Service (QoS). Static routing
protocols don’t give efficiently achievements when it
works in the dynamic environment of MANET networks.
So, the requirement of dynamic routing protocols is
needed. The key purpose is finding the better route which
can provide the various requirement of QoS constraint.
This analysis relies on distinct elements such as PDR
(packet delivery ratio), the throughput of data sent and
E2E (end-to-end) delay to reach the best protocol that can
be used if the network is low density [2].To achieve this
purpose the fruit fly algorithm (FOA) was proposed and it
was used to simulate in Mat lab program, the simulation
results obtained for classical AODV, PSO and the
proposed FOA methods and by analyzing and comparing
these protocols at different low -density nodes it was
found that the FOA protocol is the best among the
protocols studied within the conditions specified in this
study.
II. LITERATURE REVIEW
Several researchers have investigated algorithms for
routing protocols in MANET to find the suitable route
between nodes of wireless linked. The authors in [3] study
the performance and efficiency of MANET Network and
it has been improved by integrating the heuristic approach
with AODV. The simulated was under a different node
condition. The work yields better results compared to the
ACO and GA models. In [4] the authors proposed a
genetic algorithm that finds the optimal path between
nodes in MANET, using NS2 network simulation and the
NAM tool based on two different routing methods
conventional AODV and GA that yielded better results.
In [5] a comparative study between DSDV, AODV
and DSR is achieved. The study focused on proposing an
extension of classic routing protocols that would be
preferable in the form of safety, the productivity of
network, the effective utilized of restricted purses, and the
service quality. In the study of [6], Another interesting
article is to be proposed where the authors tested and
compared the multiple routing of Ad hoc protocols that
included the OLSR protocol, the DSDV protocol, the
AODV protocol, DSR, protocol ZRP protocol and
TORA protocol, which were capable to work in an
efficient manner. Then the study of [7] included the
classified of the MANET routing protocols as proactive,
interactive, and hybrid protocols. The study of the
simulation was done to find a comparison between the
rendering of two routing protocols DSR protocol and
AODV protocol on demand and the single routing
protocol DSDV gave by utilizing a performance table
with different parameters.
In this article, the approach of Fruit Fly was proposed
and accomplished to deal with the case of optimized
routing for the MANET network from the deliver node
toward the node of destination. It included study the
influence of the nodes of mobile and its random
movement in the MANET network. The performance
comparison of the proposed protocols such as PSO-
AODV and traditional AODV would be made on the base
of quantitative measures, which included the throughput
of data packets, (PDR) packet delivery ratio and the delay
of time.
III. ROUTING
The Internet is a very large network that connects to
the entire world so that it is not possible to know the
location of all the links to all computers connected. The
routing specifies the communication path from the
message transmitting station to the receiving station. In an
ad hoc wireless network, each node may move or be
unable to connect, and the connection path may suddenly
be interrupted. It is necessary to design a routing protocol
that can handle these dynamic changes [8]. Computer
networks use the information to arrange data destinations
(packets) that are continuously transmitted. It is called the
routing table. It is utilized to record the routes for each
node, and it includes the transfer destinations that are used
to send the deliver packets to destinations nodes. If the
data packet reaches its target node, it will receive it while
it is directed to itself, but with the routing table is routes
and it refers to the adjacent nodes for the destination node.
This is called routing as illustrated in Figure 1. To send
the package from node A to node E, utilizing the table of
routing for the node A, the node E will send to destination
node D, so that the package is sent to node D. then in the
routing table for the node D, it is included a destination E
which will be sent to node E, so the package is received in
node E because the destination is the same. So, the packet
will successfully receive to the destination node E. This is
the "routing protocol" role. In the MANET network, the
nodes move frequently, the links may continue and stop,
and the nodes may stop suddenly. Since the protocol of
optimal routing changes depending on the different
variations of environment included the characteristics of
node movement and the frequency of nodes.
MANET (Mobile Ad hoc NET work)
IV. AD HOC ON-DEMAND DISTANCE VECTOR
(AODV)
It is one of MANET’s most popular routing algorithms
and it is an interactive type routing protocol that only
specifies paths upon request. The basic strategy is as
follows [10]: The node that wants to send a transmission
request is transferred to all neighboring nodes. The node
receiving the transfer request redirects it one by one and
eventually arrived the target node. The target node sends a
response message in a reverse order to which the transfer
request message was transferred, and a communication
path has been created from the delivered nodes and the
received nodes. Since each node has the sequence number
and actively uses it for routing, each node contains
information about the next transmission location and a
valid routing table for a very short period and the package
are forwarded using it, and each routing table entry has a
foreground list that is used when a link failure occurs [7].
When a road to a new destination is needed, RREQ a
Route Request Message is created to the network to
discover the route of packets. Finally, when the RREQ
message reaches the target transmission destination, the
sending destination node returns RREP a Route Reply
message is the return to the transmission node by unicast.
As a result of this exchange, a two-way path to the source
and destination is created on the node path table in the
middle, after which, the data can be sent and received
using this routing table.
V. QUALITY OF SERVICE
QoS (Quality of Service) is a desired quality for
MANET as the result of the development of multimedia
applications. A process for directing traffic from network
devices, such as routers and adapters, to the behavior
required by an application that creates traffic. In other
words, QoS allows network devices to distinguish traffic
and then apply different measures to traffic [9]. It is
necessary for traffic management in package based
networks. QoS is a collective influence of service
characteristics which provide the degree of contentment of
the services of the users [10]. Features of QoS are types of
requirements services to be satisfied by the network by
sending the data packet from delivering node towards the
destination node. Multiple QoS parameters were
determined and observed to indicate if the requested or
received a level of service is achieved. The routing of QoS
is efficient for MANET and needs not only to implement a
route from a deliver node toward the destination node of
the destination but furthermore the route may achieve the
E2E delay of QoS demands. For achieve QoS (the
bandwidth and delay warranty), the stream of the packet
(route detection) and the structure of the routing table is
revised to implement the demands of service that should
be achieved by the data nodes with forwarding RREQ and
RREP package[11].
VI. THE PSO APPROACH
Particle swarm optimization approach was defined as a
population depended on the computational technique. It is
depended on the behaviour of flocks in nature such as bird
swarm and fish flocks that was proposed in 1995 by
Kennedy and Eberhart [12]. It is used in different
applications to find the required parameters for
minimizing or maximizing. In each iteration, the posi and
spin vector of a particle is updated in each dimension i
utilizing (1) and (2). Each particle of i represented by the
vector posi. The fitness function is utilized to find the
activity for each particle. Each particle will fly depend on
the experience and from the social environment and the
current position and the particle. The experience of the
particle i represented as ptbest of the best position found
Fig.1. Routing illustration.
by these particles[13]. Each member has a memory that
saved the best previously searching locations with the
fitness function in that position that is updated through
time. The information taken from the environment was
represented by particles that have the best position gt in
the flock of the particles, with the current position of
particle i is expressed by posi(t). The first step in PSO is
initialization includes the number of iterations, the number
of population n and the inertia weight W [14]. The
coefficients of acceleration c1 and c2 are cognitive
learning which two positive constants are. Moreover, the
next step has utilized the population in the form of a
random matrix with a range [0,1] of values. After that, the
initialization of speed and position is achieved in this step,
so that we will make the values of the velocity and
position for each particle equal to zero, and then calculate
the error. To determine the variation into speed and
position of the particle (spi, posi) the following equations
are utilized as indicated below [13]:
sp
𝒊
(t+1) = W*sp
𝒊
(t) + µ1 *c1 (ptBest − posi(t ))
+ µ2 * c
𝟐
(gtBest− pos
𝒊
(t)) (1)
with pos is pos
𝒊
(t+1) = pos
𝒊
(t) + sp
𝒊
(t +1) (2)
The variable µ vector is a random value with a range
value [0,1]. The speed of each particle is limited between
[spmin, spmax]. W is the inertia weight provides the
balance between global exploration and local exploration
of PSO. posi and spi is the current position and velocity
of a particle. gtBest represents the global best position of
the best particle element. PSO needs a large number of
iterations in searching for the optimum solution, W
approximately ranging from 0.9 to 0.4 for the calculations.
General inertia weight W is set below [14].
W = W
𝒎𝒂𝒙
– ((W
𝒎𝒂𝒙
− WWIN )/ ITER
𝒎𝒂𝒙
)* ITERNO (3)
The PSO approach work as shown in Fig. the particles
elements are distributed randomly for each initial position
in the environment of the workspace at first, then the
value of velocity of each particle is randomly assigned
and it will continue updating itself depends on the
particle’s own and adjacent experience or the expert of the
entire flock[14].
VII. THE FAO APPROACH
Fruit Fly is an evolutionary computation approach, it
was first proposed by Wen T. Pan in 2011[15]. Fruit fly
flock in natural includes the behaviour of social which
utilizing the intelligence of the collective to achieve bases
activities. FOA was depended on the behaviour of food
searching of the fruit fly flock [16]. FOA utilizes
osphresis and vision superior to other species. Fruit flies
capable to smell the food source from far away distance
and then fly fast toward that direction. The fruit fly has
two phases for food research process [16, 17]:
1. In the first phase, the fruit fly utilizes the sense of
smell capability to fly through the food area to search the
food source then moves to the specified direction.
2. The vision capability of flies is utilized to get
closer in the second phase after it approaches the location
of food. When flies become near the source of food, the
vision sensitive of the flies utilized to locate food position
then flying in that direction.
The fruit flies have a superior smell and seeing
capabilities compared with other flies sorts. The flies are
able to smell the food even for a distance of 40 km away
[16]. The food searching behaviour of fruit flies was
indicating in Fig.3. The steps of the FOA algorithm
flowchart are given in fig.4 to show its action. The
algorithm steps can be shown that the smell phase of the
algorithm is short compared with the vision phase.
The food-finding steps of FOA [17] are as indicated
below:
1. Adjust the parameters; a flock of n flies is located
initially into the source node Si.
2. Locate the fruit flies in random positions.
Fig.3. Food searching of fruit flies.
Fig.2. the PSO structure
3. Each fruit fly will move from the source node towards
the neighbour location of next node randomly; the
searching operation to the source of food by applying (4)
xpi= xaxis + rand (4)
ypi = yaxis + rand value
4. The value of fruit fly smell concentration was
calculated as shown below:
(5)
Spi=1/Disi , smll =function (Si)
5. To find the location of fly that gives the best smell
concentration value and the value of best position of the
flock in that location, the next iterations is found utilizing
(6)
[Best smll Best Indx]=MAX (smll) (6)
xaxis = xp(Best Indx) , yaxis = yp(Best Indx)
6. The iterations are repeated until the fruit fly discovers
the food location for the destination node.
7. When the number of maximum iteration is reached, the
FOA algorithm method will report the best solution value,
otherwise, it will return back to step number 3.
The distribution mechanism and development of the
individual for the flock of fruit flies has been improved to
increase research speed and accuracy [17].
VIII. SIMULATION ANALYSIS
To apply the suggested routing protocol, we prepare a
network workspace of the environment first, after that we
implement and make a comparison between FOA
proposed approach with the multiple scenarios. The
simulation results in this study were done utilizing Mat
lab program. The network simulation having two parts as
indicated below:
1. Performing the MANET network utilizing the
protocol of AODV: In that scenario of the network, it was
achieved a network using AODV as a routing protocol
with different nodes number in the network.
2. In this part, the network protocol was achieved
utilizing the traditional AODV protocol and supported
with optimized approach PSO and FOA algorithms.
IIIV. PERFORMANCE METRICS
This section of our simulation study gives the analysis
and metrics of the suggested routing concerning
traditional AODV as routing protocol. For performance
analysis, we utilized different parameters, different
scenarios and metrics. The metrics elements which are
utilizing in our study are shown in Table I [4].
A. The packet delivery ratio
The ratio between the data delivered to the node of the
destination to the data transmitted packets generated by
the deliver source node [18]. PDR Packet Delivery Ratio
gives the necessary information required about the
performance of any routing protocols. Where PDR is
evaluated utilizing the following equation:
𝑃𝐷𝑅= (delivered of data packets)/ (sent data packet)∗100 (7)
The comparison of packet delivery ratio with a
different number of nodes is indicated in fig.5.
Fig.5. Comparison of no. of nodes and packet delivery ratio
PDR%
TABLE I. SIMULATION PARAMETERS
Parameters
Description
Number of nodes
5,10, 20,40, 60
The type of channel
Channel is wireless
The type of network
interface
Wireless Phy
Mac type
Mac/802_11
Simulation time
60 seconds
protocol of Routing
AODV protocol
The size of Simulation
1000 x 1000
Fig.4. FOA algorithm
B. E2E Time delay
The term E2E refers to the average time of the end to
end delay of the data packets which is needed to transmit
the data packet from delivering node to the node of
destination [19]. The lower is the end to end delay, the
better the application performs. This metric indicates the
time duration for the transferring of the data packet from
delivering a node to the node of destination. Total time
difference between send and receiving of any data.
E2E delay = (8)
The E2E time delay comparison is indicated in fig.6.
C. Throughput (TP) data
This metric is equal to the number of bytes which the
destination has received.
TP = ( (9)
Throughput data provides the efficiency of the system,
after evaluation of the performance parameters, we can
see that the overall performance of the proposed FOA-
AODV algorithm is much efficient than the traditional
routing protocol AODV and PSO-AODV as shown in
Fig.7. The FOA algorithm is faster and stable approach
compared with other methods in solving the optimized
problems, it is also an easy in implementation and
application for the optimized problems.
IX. FOA ALGORITHM WITH DIFFERENT NETWORK
TOPOLOGIES
In this section, the routing problem was optimized by
using the fruit fly approach with multiple network
topologies [18], by adjusting the nodes number, the edges
and the configurations of links. The simulations
experiments are implemented in different topologies are
named problem cases a, b, c, d and e. Each case provides a
topology contains a different nodes number and different
numbers of links as indicated on table II. Fig. 8 indicates
the network topology of case .a with five nodes. Then, the
proposed approach of FOA-AODV was applied to get the
routing objective function, and the complexity time and
Fruit flies numbers: 5, 10, 20, 40 and 60 as indicated in
Fig.9 and Fig.10.
Moreover, by applying the proposed FOA-AODV
approach, it was found that the ability to achieve the
optimal solution to the problem will increase if the size of
the flock increases. It was observed that the minimum
routing objective function results are similar in the small
size network regardless of the size of the flock as
indicated in Fig.9.That implies the FOA-AODV
optimization approach can get the optimal solution in the
network of small size by utilizing a small flock size. The
average complexity time as a function of five cases of the
proposed fruit fly optimization approach is indicated in
fig.10. The ACT (Average Complexity Time) of the FOA-
AODV approach will be increased when the number of
the flock size increased for all problems. The ACT for the
routing problem is founded by adjusting the number of
nodes and links in the experiments, and in terms of the
comparison results indicated that the FOA-AODV
overcome the classical method in terms of ACT and it is
efficient for the routing optimization problem.
Fig.8. Network topology of case a with five nodes
Fig .6. E2E delay time of the three protocols
Fig.7. The Throughput of data (TP)
Fig.7. the throughput of data (TP)
TABLE II. THE PROBLEMS FOR THE CASES
Cases
No. of nodes
No. of edges
Case a
5
8
Case b
10
22
Case c
20
99
Case d
40
158
Case e
60
302
X. CONCLUSION
The application of FOA optimization approach was
achieved in this study, which is the typical meta-heuristic
optimization method. The fruit fly optimization approach
is a global optimization algorithm depended on the
behavior of the fruit fly flock to foraging the food. The
suggested optimized approach was applied effectively for
routing to discover the optimal shortest route path
between the source nodes toward the destination node.
The proposed scheme for FOA-AODV algorithm to get
the shortest optimal route shows efficient results
compared with traditional AODV and PSO-AODV. The
performance of the FOA-AODV algorithm was
determined by implementing the experiments with distinct
numbers of nodes and links, and by applied ACT in the
MANET routing network, with classic test functions,
simulation improvement results show that the proposed
FOA algorithm offers the fastest and most accurate
pathway and more reliable optimization ability. The
compared results indicated that the FOA-AOD approach
overcomes the conventional methods in the average
complexity time state and it is adequate for solving the
routing optimization problem.
REFERENCES
[1] N. Raza, M. U. Aftab, M. Q. Akbar, O. Ashraf, and M. Irfan,
“Mobile Ad-Hoc Networks Applications and Its Challenges”,
Communications and Network, 2016, 8, 131-136, 2016.
[2] M. E. Manaa, and S. D. Shamsi, “Improved MANET Routing
Protocols Performance using Optimization Methods”, International
School of Energy Science and Engineering, Harbin Institute of
Technology, Harbin, P. R. China,2014 .
[3] N. Jagadev, and B. K. Pattanayak, “Power aware routing for
MANET using PSO”, International Journal of Innovative Technology
and Exploring Engineering (IJITEE) ISSN: 2278-3075, Volume-8 Issue-
6, April, 2019.
[4] H. Lafta, A. Al-Salih ,” Efficient Routing Protocol in the Mobile Ad-
hoc Network (MANET) by using Genetic Algorithm (GA)” ، IOSR
Journal of Computer Engineering (IOSR-JCE) e-ISSN: 2278-0661, p-
ISSN:2278-8727VolumeX,IssueX,Ver.I,pp.01-00, Jan,2014.
[5] C. Wahi, and S. K. Sonbhadra, “Mobile Ad Hoc Network Routing
Protocols: Comparative Study” , International Journal of Ad hoc, Sensor
& Ubiquitous Computing (IJASUC) Vol.3, No.2, April, 2012.
[6] R.Ismail, C. Z .Zulkifli, and K.Samsudin,”Routing protocols for
mobile Ad-Ho implementation and application of the problems when
compared with other algorithms network: A qualitative comparative
analysis”, Journal Technology (Sciences & Engineering), 2016.
[7] N. Kohila and R. Gowthami, “Routing Protocols in Mobile Ad-Hoc
Network”, Vivekananda College of Arts and Sciences for women,
Namakwa, Tamil Nadu, India, International Journal of Computer
Science and Mobile Computing A Monthly Journal of Computer Science
and Information Technology ISSN 2320–088X IJCSMC, Vol. 4, Issue.
1, pp.159 – 167, January 2015.
[8] K. Gorantala, “Routing Protocols in Mobile Ad-hoc Networks”,
2006.
[9] S. Liu Yang and W. Wang, “Research of AODV Routing Protocol
for Ad Hoc Networks”, School of Information Engineering, Chang’an
University, Xi’an, 710064, 2013.
[10] M. Varela, L. S. Kapov and T. Ebrahimi, “Quality of Service vs.
Quality of Experience”, 2014.
[11] R. S. Camama, A. L. Alcaraz , S. B. Angot, and J. Claveria,
“Quality of Service (QoS) in School Networks”, School of Computing
and Information Technologies, Asia Pacific College Makati, Philippines,
2016.
[12] E. Russell, and J. Kennedy, “A new optimizer using particle swarm
theory”. In Micro Machine and Human Science, MHS 95, Proceedings
of the Sixth International Symposium on, pp. 39-43. IEEE, 1995.
[13] A.T. Sadiq, F. A. Raheem, N. Alhuda, and F. Abbas,” Robot Arm
Path Planning Using Modified Particle Swarm Optimization based on D*
algorithm”, AL-Khwarizmi Engineering Journal, Vol. 13, No. 3, P.P. 27-
37, 2017.
[14] M. F. Aranza, J. Kustija, B. Trisno, and D. L. Hakim, “Tunning PID
controller using particle swarm optimization algorithm on automatic
voltage regulator system”, International Conference on Innovation in
Engineering and Vocational Education, Materials Science and
Engineering 128, 2016.
[15] P. WA “new fruit fly optimization algorithm taking the financial
distress model as an example”. Knowl.Based Syst. 26(2), pp 69-74,
2012.
[16] X. Li, L. Sun, J. Li, and H. Piao, “An Improved Fruit Fly
Optimization Algorithm and Its Application in Heat Exchange Fouling
Ultrasonic Detection”, mathematical problems in engineering journal,
2018.
[17] H. Iscana, and M. Gunduz, “An application of fruit fly optimization
algorithm for traveling salesman problem”, International Conference on
Advances in Information Technology, IAIT2016, 19-22,2014.
[18] Z. He, H. Qi, Y. Yao and L. Ruan,” Inverse estimation of the
particle size distribution using the Fruit Fly Optimization Algorithm “,
[19] S.M. Darwish, A. Elmasry, and S.H. Ibrahim, “Optimal shortest
path in mobil Ad-Hoc Network based on Fruit Fly Optimization
Algorithm”, Springer Nature Switzerland AG, AMLTA, AISC 921, PP
91-101,2019.
[20] M. H. Hassan, S. A. Mostafa, M. A. Mohammed, D. A. Ibrahim, B.
A. Khalaf, and A.S. Al-Khalifa,” Integrating African Buffalo
Optimization Algorithm in AODV Routing Protocol for Improving the
Qos of MANET”, Journal of southwest Jiao tong university, Vol.,
No.3.June, 2019.
Fig.9.the min routing objective function
Fig.10. The average complexity time of routing
