Content uploaded by Mustafa Hamid Hassan
Author content
All content in this area was uploaded by Mustafa Hamid Hassan on Jul 07, 2019
Content may be subject to copyright.
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 1 (2017) pp. 134-139
© Research India Publications. http://www.ripublication.com
134
An Improved Hybrid Technique for Energy and Delay Routing in Mobile
Ad-Hoc Networks
Mustafa Hamid Hassan1 and Ravie Chandren Muniyandi2
Center for Software Technology and Management (SOFTAM), Faculty of Information Science and Technology.
Universiti Kebangsaan Malaysia (UKM), 43600, Bangi, Selangor, Malaysia.
Abstract
A mobile ad-hoc network (MANET) comprises wireless
mobile nodes that dynamically establish a temporary network
without needing a central administration or an infrastructure.
The Quality of service (QoS) enables a mobile network to
interconnect wired or wireless networks. An important
research background identifying a path that satisfies the QoS
requirements, such as their topology and applications, has
become quite a challenge in mobile networks. The QoS
routing feature can also function in a stand-alone multi-hop
mobile network for real-time applications. A QoS-aware
protocol aims to find a stable path between the source and
destination nodes that satisfies the QoS requirements. The
proposed method was a new energy and delay-aware routing
protocol that combines cellular automata (CA) with the hybrid
genetic algorithm (GA) and African Buffalo Optimization
(ABO) to optimize the path selection in the ad-hoc on-demand
distance vector (AODV) routing protocol. The main
conclusions of this research include two QoS parameters that
were used for routing: energy and delay. The routing
algorithm based on CA was used to identify a set of routes
that can satisfy the delay constraints and then select a
reasonably good route through the hybrid algorithm. Results
of the simulation showed that the proposed approach
demonstrated a better performance than the AODV with CA
and GA.
Keywords: mobile ad-hoc network, cellar automata, genetic
algorithm, African Buffalo Optimization, ad-hoc on-demand
distance vector, quality of service.
INTRODUCTION
A mobile ad-hoc network (MANET) exhibits a dynamic
topology that does not contain any fixed infrastructure, with
each node having host and router functionalities [1]. The most
critical features of a MANET include autonomy and the
absence of infrastructure, dynamic network topology, multi-
hop routing, limited physical security, device heterogeneity,
and variable capacity links with constrained bandwidth [2,
11]. The attributes of a MANET have various applications
even with numerous constraints [15, 20]. These attributes
include high ability in circumstances where a fixed
infrastructure does not exist [19, 9]. The second feature is that
a MANET does not have to operate on its own because it can
be attached to the Internet and be incorporated in various
devices, making its respective services available to the rest of
the users. MANETs have been used in numerous applications
in the past. Establishing the path from source to destination is
crucial when using MANETs to deliver a data packet that
satisfies the quality of service (QoS) standards, such as the
end-to-end delay, throughput, and energy. Nevertheless, the
algorithm designed in this study should be comparable with
the QoS-based method in terms of the average end-to-end
delay. However, the proposed algorithm is more effective in
terms of node lifespan and packet delivery ratio. African
Buffalo Optimization (ABO) [4] constitutes an effort to
develop a convenient, reliable, efficient, effective, and easily
implementable algorithm that will exhibit outstanding ability
in exploring and exploiting the search space. The hybrid
algorithm will be integrated with the ad-hoc on-demand
distance vector (AODV) routing protocol to improve the QoS.
Numerous artificial and heuristic QoS routing algorithms have
been suggested for application in MANETs routing [1, 5].
Basically, routing is divided into several categories: single-
and multi-path routing; source routing, and next step,
hierarchical, and flat routing; centralized and distributed
routing; data- and address-centric routing; QoS-based and
best-effort routing; event-driven and queue-based routing; and
energy-based routing. Many articles have discussed this issue
and used several methods, which were often heuristic and
intelligence tools.
Genetic algorithm (GA) is one of the most popular methods.
A fuzzy GA is used for QoS routing [7]. The exact
information protection of global network status is impossible
for the nodes of a real dynamic network. Therefore, at first
QoS parameters were fuzzed, and then became fuzzed to
optimize the fitness function using the GA. Meanwhile, one of
the related studies on multiple QoS routing algorithms based
on GA was introduced [8]. Evolutionary optimization
strategy, which is used in QoS multiple routing, is one of the
other intelligent methods [17]. In the article, an evolutionary
multi-purpose quick method was proposed as the Multi-
Objective Evolutionary Algorithm (MOEAQ) to find the
optimized path of the QoS. It was better than its basic version
because of its convergence. In addition to high diversity, it
could obtain more favorable results than the popular routing
algorithm based on GA [12]. QoS routing could be performed
by comparing several intelligent methods [3].
Swarm intelligence is a new and modern field inspired by the
swarm behavior of creatures [12, 6]. The cellular model
simulates the natural evolution from the perspective of the
individual, which encodes a tentative (optimization, learning,
search) problem solution. A cellular automaton (CA) has been
shown capable of resolving various complex systems, such as
MANET and issues in varied applications. A popular
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 1 (2017) pp. 134-139
© Research India Publications. http://www.ripublication.com
135
approach based on CA and employed in routing includes the
Lee algorithm, which identifies the path that has the lowest
aggregate weight on the grid [18]. The algorithm needs to
establish the shortest path from the source cell to the
destination cell when the weights of all the nodes (grid points)
are configured as one. In the framework utilized, the authors
considered the possibility of a relationship between the
number of states and the longest path or the largest aggregate
weight. Meanwhile, GA constitutes a search heuristic that
mimics the natural process of evolution and is normally
employed in generating solutions applicable in handling
search and optimization problems [10, 13]. The composition
of an ad-hoc network includes spatially distributed
autonomous systems where intermediate nodes communicate
with each other. Every mobile node can deal with route
broadcasting even without requiring a consolidated controller.
However, the main challenge is on lowering the power
consumption of the various mobile nodes because each node
has a limited battery power.
The GA-based routing algorithm has been observed to lead to
the development of a heuristic methodology for MANETs.
Nevertheless, the algorithm tends to speedily converge to a
single solution through GA but causes network congestion.
The work in [18] attempted to integrate genetic algorithm
(GA) with cellular automata (CA) to improve efficiency.
However, the main objective of this study is to solve the
following disadvantages:
a. GA does not guarantee that the global maxima will be
attained. However, in addition to brute force, GA does
not guarantee the global maxima when dealing with non-
trivial issues. However, the possibility of being fixed to
the local maxima at initial phases is an issue that must be
addressed, for instance, with a form of simulated
galvanizing mutation rate decay.
b. GA consumer time before convergence is reached. A
decently sized population and numerous generations are
necessary to attain desirable results. With heavy
simulations, a solution will often take days to obtain.
AN EFFICIENT ENERGY AND DELAY ROUTING
PROTOCOL
Hybrid Algorithm Gaabo
The genetic Algorithm have been used to solve different
problems, one of them is an energy problem that shown in
Ahmadi scheme. But the GA still consume a high amount of
time to get an optimum solution moreover, stuck in local
optima. This research proposed a hybrid algorithm to solve
the GA problems. This algorithm combines the advantages of
both GA and ABO algorithms. The ABO is a response to
create the population and discover the initial solution. The GA
selects two solutions from an initial solution that comes from
ABO. Then, it performs the crossover and mutation process
till produce the best solution for GA algorithm. Next, ABO
checks the state of the solution based on the fitness.
Thereafter, it will make an update for the buffalos’ position.
Then check the stopping criteria. Subsequently, get an
optimum solution. Overall steps for a hybrid algorithm are
presented below.
Step 1: Initialization of Population is done randomly.
Buffaloes are placed at nodes in a random manner.
Step 2: Determine Fitness of population by the following
equation
(1)
where
Lp1,lp2 learning factors
Bgmax herd’s best fitness
Bpmax.k individual buffalo’s best fitness
Wk mk exploration and exploitation movement of kth
buffalo
K=1, 2…., N.
Step 3: Repeat
o Select parents from the population.
o Perform Crossover on parents, creating a new population.
o Perform Mutation on New population.
o Determine Fitness of the population.
o Update location of New child (bgmax, bpmax) by using the
below equation:
(2)
o Check bgmax is getting updated, If Yes Check Stopping
Criteria (Step 4) If No Initialize population (go back to
step 1).
o Check Stopping Criteria, if met go to Step 4 else go to
step 2.
Step 4: Output Best Solution.
All the proposed stages can show in figure 1.
Figure 1: hybrid algorithm
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 1 (2017) pp. 134-139
© Research India Publications. http://www.ripublication.com
136
Implementation of Discovering Routes by CA
The aspects of the executions are understood from the
explanation of the mapping of CA using the MANET
topology. A two-dimensional network plane in which several
mobile nodes are spread randomly is considered in this study,
to find out a path with delay as the QoS constraint. There
exists no hierarchy among the nodes, and the network plane is
found to be fully homogeneous (i.e. all nodes consist of the
same characteristics).
Our approach involves the broadcasting of the RREQ message
that constitutes the delay requirement of the connection
request [maximum delay (Dmax)] by the source node to its
communicating neighbors. All the nodes at the right, left, top,
and down side are involved in this, as depicted in Figure 2.
Figure 2: CA mechanism
The message re-broadcast by the intermediate nodes to their
neighbors, which also establish a reverse path to the sender.
Certain nodes, when given a delay constraint, turn into a
wave, take in a wave node to their neighbors, re-broadcast the
message, and establish a reverse path to the nodes from which
they had obtained the message. This activity continues till the
message is collected by the destination node or the delay
faced by the packet outstrips the limit Dmax. The destination
obtains many RREQ messages for the same sender when there
are more paths from the sender to the destination.
Consequently, reply to some of the RREQ messages is done
by the destination through sending an RREP message through
the reverse path that is established when the RREQ messages
are passed on. The entire set of nodes observed along these
routes amidst the source and the destination constitute the path
nodes. Each and every communication between the source and
the destination from this juncture happens through this path
till the topology of the network gets modified... However, the
Pseudo code for the MATLAB developed code can found
bellow:
Algorithm for two-way dimensional of CA for the initial and
shortest path.
i. Input the radius of the CA K=1;
ii. Put loop for the loop for discovering and checking loop
for checking both side route from sn-dn and dn-sn
iii. checking the neighbours of CN node ,which [r,c]=
find(O == CA)
iv. storing path created by nag node with concatenating
original sn node
v. concatenating dn to created path
Figure 3 shows the rule process of selecting a path by CA.
Recursive
update rule
Selected route
for each request
suitable paths will considered
Initialize network
Output (CA)
Figure 3: The Process rule for short path selection
System Model
In general, the reactive routing protocols create a path by
using two types of messages; route request and route reply.
The source node broadcast the RREQ to discover the possible
path to the destination when the destination node receives the
RREQ message it will send an RREP message to the source
node. In protocols like AODV, the path is creating via sending
an RREQ by source node till receiving an RREP from a
specific destination. However, the AODV select the path
based on the first RREP that comes from destination rather
than the energy level of the nodes, this mechanism leads to
exhaust node in the path and increase the probability of link
failure, the link failure has negative effective on the QoS of
the network such as; throughput, E2E delay, energy
consumption, etc.
This paper, attempt to establish a robust path and get fulfill
the QoS requirement as energy and delay. This research
comprised two stages to achieve our objective, in the first
stage we used CA to discover all possible paths based on
minimum time, the second stage selects the path based on
highest energy level for each node in the path by using hybrid
algorithm GAABO. We proposed hybrid techniques that will
enable the discovery of routes in MANETs which satisfy both
the delay constraints and some simple energy constraints
(every node on a path has a minimum energy level).
Aforementioned, the CA generate the paths from source to
destination nodes based on the minimum delay, the RREQ
message that sends by CA content a threshold term to ensure
all paths achieve the requirement of delay. Subsequently, the
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 1 (2017) pp. 134-139
© Research India Publications. http://www.ripublication.com
137
hybrid algorithm checks the status of paths based on energy
level. A hybrid GAABO with CA shown in figure 4.
Initialize
Network
Output (CA)
Select path list in CA
(ABO initialize population )
ABO feed the fitness of GA
Fitness Evaluation
Selection
Genetic operations
End
Maximum Location
Recursive
update rule
Crossover Mutation
No
Yes
Figure 4: The process of CA with hybrid algorithm
The behavior of the ABO algorithm in nature is searching for
the best pasture. The herd of buffalos’ discovers the best place
by sending different buffalo to a different place in the space.
The buffalos’ communicate between each other by sound.
Each buffalo has two kinds of sound (maaa, waaa). The maaa
sound means a place is a good place for the herd. On the other
hand, the waaa sound means the place is not good (no pasture
or danger) for the herd.
These characteristics of ABO are used in MANET. Figure
(4.4) illustrate the steps of proposed approach. As mentioned
before, the CA creates an initial population based on less
delay. Then, the ABO use the population that generated by
CA to select the path based on energy, each node in the
network represent as buffalo in ABO, as mentioned earlier the
buffalo (node) has two types of sounds( messages )these
messages are uses to indicate the energy level of each node in
the networks. All nodes forward the information of energy to
the source node (herd). There are different paths are discover
from source to destination nodes. The fitness function of ABO
evaluates the quality of paths then sorts them. Moreover, get
an optimum solution from all possible solutions, the GA
selects two best solutions from a list of solution that generated
by ABO as parents. Thereafter perform the operation of GA
(crossover and mutation), then select the best solution (path).
Through this mechanism get an optimum path with QoS
requirements (less delay and highest energy level).
SIMULATION ENVIRONMENT
This section represents the parameters that use to implement
the proposed approach and AODV. MATLAB toolbox was
utilized to implement and evaluate the performance of routing
protocols. To evaluate the behavior of protocols we changed
the nodes speed as (5, 10, 15, 20 and 25), with node density
625 node, these node distributed in Random Way Point
(RWP) within 3000m2, packet size 512, CPR is controlled the
traffic and pause time 5ms. The different performance metric
utilized to evaluate the performance of the protocols:
a. Packet loss ratio (PLR): define the ratio of the
difference between the number of data packets sent
and received.
b. Packet delivery ratio (PDR): the ratio of the data
delivered to the destination node to the packets
transmitted by the source.
c. End-to-end (E2E) delay: This metric represent the
amount of time spent to transmit the data packet from
source to destination.
d. Throughput (TP): defined as the number of bytes
that have been successfully received by the
destination.
e. Energy Consumption: defined as the amount of
energy consumed by all nodes in the network in a
given simulation time.
RESULTS AND DISCUSSION
After the simulation was completed for the proposed method,
performance analysis was conducted using the evaluation
metrics, PDR, E2E delay, energy consumption, PLR, and TP.
A simulation study was conducted by varying node speeds of
5, 10, 15, 20, and 25 m/s to show the effectiveness of energy
cost on AODV routing protocol using CA with hybrid
GA+ABO.
Figure 5: Packet Delivery Ratio
Figure 5 shows the packet delivery ratio for protocols, which
have a different PDR. However, the proposed approach has a
better result in PDR than CAGA protocol. Due to the CA with
GAABO establish a stable path based on less delay and
highest energy between source and destination nodes. This
minimizes the probability of link failure as well as packets
loss.
86
88
90
92
94
96
98
100
510 15 20 25
PDR (%)
Mobility
CA+GA
CA+GAABO
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 1 (2017) pp. 134-139
© Research India Publications. http://www.ripublication.com
138
Figure 6: E2E delay
Figure 6 depicts the E2E delay for CAGA and CA with
GAABO. That has a different E2E delay with the increase of
the node mobility in the network. In the term of E2E delay,
the proposed approach is better than CAGA protocol. Since
the CA with GAABO searching for the path from source to
destination nodes relay on the minimum delay between the
nodes.
Figure 7: Energy Consumption
The variation of energy consumption for the CA with
GAABO and CAGA are presented in figure 7. For all
strategies, as the node mobility increases the energy
consumption increases also. Results clearly show that
proposed approach is better than another approach in term of
energy consumption. Because once the protocol selects the
best path, then, the same path will be used to transmit all
packets. This path is highest energy level. Therefore, as the
intermediate nodes will not perform route discovery and it
need not have to waste its battery power.
Figure 8: Packet loss
Figure 8 shows the variation of packet loss with mobility.
When the mobility increases as (5, 10, 15, 20, 25) m/s, the
packet loss increases also. The CA with GAABO has better
performance than CAGA protocol in term of packet loss. Due
to proposed protocol selects the route with less delay and
highest energy, which saves time for the packets to be
transmitted over the network.
Figure 9: Throughput
The variation of average throughput for protocols is shown in
figure 9. While the node mobility increased as (5, 10, 15, 20,
25) the throughput decrease in routing protocols. CA with
GAABO protocol has better throughput than CAGA, as it
selects the most active route to the destination. This route has
less delay and more energy level than other routes; therefore
the link is more stable which leads for fewer packets dropped.
This, in turn, increases the throughput.
CONCLUSION
A wide range of fields, such as commercial and military
applications, employ MANETs. Thus, establishing a path
from source to target is essential to ensure that the data packet
delivered meets the QoS requirements. However, this paper
proposed a QoS-routing algorithm applicable in MANETs,
which satisfies energy and delay constraints. In other words,
the algorithm must support the QoS parameters of energy and
0.1
0.2
0.3
0.4
0.5
0.6
510 15 20 25
E2E Delay
Mobility
CA+GA
CA+GAABO
6
8
10
12
14
16
510 15 20 25
Energy Consumption
Mobility
CA+GA
CA+GAABO
0
10
20
30
40
510 15 20 25
Packet Loss
Mobility
CA+GA
CA+GAABO
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
510 15 20 25
Throughput
Mobility
CA+GA
CA+GAABO
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 1 (2017) pp. 134-139
© Research India Publications. http://www.ripublication.com
139
delay. Using CA with GAABO techniques, this study sought
to enhance the network lifetime as well as the E2E delay.
ACKNOWLEDGEMENT
This work is supported by the Fundamental Research Grant
Scheme of the Ministry of Higher Education (Malaysia; Grant
code: FGRS/1/2015/ICT04/UKM/02/3).
REFERENCES
[1] Conti M, Giordano S. Mobile ad hoc networking:
milestones, challenges, and new research directions.
Communications Magazine, IEEE. 2014;52(1):85-
96.
[2] Basagni S, Conti M, Giordano S, Stojmenovic I.
Mobile Ad Hoc networking: the cutting edge
directions: John Wiley & Sons; 2013.
[3] Gulati MK, Kumar K, editors. A review of QoS
routing protocols in MANETs. Computer
Communication and Informatics (ICCCI), 2013
International Conference on; 2013: IEEE.
[4] Odili JB, Kahar MNM, Anwar S. African Buffalo
Optimization: A Swarm-Intelligence Technique.
Procedia Computer Science. 2015;76:443-8.
[5] Fahad, A. M. & Muniyandi, R. C. 2016. Research
Article Harmony Search Algorithm to Prevent
Malicious Nodes in Mobile Ad Hoc Networks
(MANETs).
[6] Nivetha S, Asokan R, Senthilkumaran N. ACO-GA
Hybrid Meta-Heuristic (AGHM) Optimization for
Multi-constrained Quality of Service Routing in
MANETs. 2014.
[7] Ayyasamy A, Venkatachalapathy K. Context aware
adaptive fuzzy based QoS routing scheme for
streaming services over MANETs. Wireless
Networks. 2015;21(2):421-30.
[8] Allakany AM, Mahmoud TM, Okamura K, Girgis
MR. Multiple constraints QoS multicast routing
optimization algorithm based on Genetic Tabu
Search Algorithm. Advances in Computer Science:
an International Journal (ACSIJ). 2015;4(3).
[9] Li, Y., Li, X., Liu, Q., & Liu, Z. (2009, June). E-
PROPHET: a novel routing protocol for
intermittently connected wireless networks.
In Proceedings of the 2009 International Conference
on Wireless Communications and Mobile
Computing: Connecting the World Wirelessly (pp.
452-456). ACM.
[10] Karthikeyan P, Baskar S. Genetic algorithm with
ensemble of immigrant strategies for multicast
routing in Ad hoc networks. Soft Computing.
2015;19(2):489-98.
[11] Karthikeyan P, Amudhavel J, Infanta E, Deepika P,
Saranya M, Nandhini C, editors. Impact on Self-
Organization in Mobile Adhoc Networks: An
Comprehensive Review. Proceedings of the 2015
International Conference on Advanced Research in
Computer Science Engineering & Technology
(ICARCSET 2015); 2015: ACM.
[12] Kiani F, Amiri E, Zamani M, Khodadadi T, Manaf
AA. Efficient intelligent energy routing protocol in
wireless sensor networks. International Journal of
Distributed Sensor Networks. 2015;2015:15.
[13] Li, Y., Yu, J., Wang, C., Liu, Q., Cao, B., &
Daneshmand, M. (2012, December). A novel
bargaining based incentive protocol for opportunistic
networks. In Global Communications Conference
(GLOBECOM), 2012 IEEE (pp. 5285-5289). IEEE.
[14] Falk W, Williams J. Deformed Cellular Automata.
2015.
[15] Liu F-H, Lo H-F, Juan SC, Lee W-T, Liao J-c.
Energy-efficient ant-colony-based routing algorithm
for the MANETs. J Internet Technol. 2013;14(1):21-
30.
[16] Lu T, Zhu J. Genetic algorithm for energy-efficient
QoS multicast routing. Communications Letters,
IEEE. 2013;17(1):31-4.
[17] Spears WM. Evolutionary algorithms: the role of
mutation and recombination: Springer Science &
Business Media; 2013.
[18] Ahmadi M, Shojafar M, Khademzadeh A, Badie K,
Tavoli R. A Hybrid Algorithm for Preserving Energy
and Delay Routing in Mobile Ad-Hoc Networks.
Wireless Personal Communications.
2015;85(4):2485-505.
[19] Bang AO, Ramteke L. MANET: History, Challenges
And Applications. International Journal of
Application or Innovation in Engineering &
Management (IJAIEM). 2013;2(9):249-51.
[20] Jubair, M. & Muniyandi, R. 2016. Ns2 Simulator
to Evaluate the Effective of Nodes Number and
Simulation Time on the Reactive Routing Protocols
in MANET. International Journal of Applied
Engineering Research 11(23): 11394-11399.
