You are on page 1of 6

International Journal of Computer Science and Information Security (IJCSIS),

Vol. 14, No. 10, October 2016

Enhanced Routing Algorithm in Mobile Ad-


Hoc Network using Hybrid Algorithm
T. Prince, Dr S. Thabasu Kannan

Abstract - In this paper, we suggest a routing Routing algorithms produce attainable courses between
optimization algorithm to proficiently decide an ideal source node and destination node to accomplish the
path from a source to a destination in mobile ad-hoc base cost. Routing in conjunction with blockage
networks. To decide an ideal path for the nodes is control and admission control characterizes the
critical for transmitting information between hubs in
execution of the network. Routing algorithm ought to
thickly conveyed networks. Keeping in mind the end
goal to proficiently transmit information to its have bland target of routing methodology to be both
destination, the proper routing algorithms must be progressively reconfigurable and in view of locally
executed in mobile ad-hoc networks. The proposed accessible data.
algorithm is composed by utilizing a Tabu search with The reason for this proposed work is to create
genetic algorithm system that is an agent meta-heuristic arrangement utilizing genetic and advance the
algorithm. The proposed Tabu search with genetic arrangement utilizing Tabu search algorithms. Genetic
algorithm does two neighborhood creating operations algorithm is one in which the populace connected with
keeping in mind the end goal to decide an ideal path and every node co-advance to tackle the issue as entirety. A
minimize algorithm execution time. We contrast the
routing algorithm in mobile ad-hoc networks requires a
proposed Tabu search and genetic algorithm with
different algorithms, regarding the routing cost and period limitation administration to decide a path from a
algorithm execution time. The correlation results source to a destination since the topologies of mobile
demonstrate that the proposed Tabu search with genetic ad-hoc networks are more as frequently as possible
algorithm beats alternate algorithms and that it is changed than those of different sorts of networks. In
reasonable for adapting the routing optimization issue. the event that the routing algorithm can't give a period
limitation administration, then the nodes can't
Keywords: appropriately transmit information to the destinations
Mobile Ad-Hoc Network (MANET), Tabu search, on time. In this way, keeping in mind the end goal to
Genetic Algorithm (GA), Routing.
give time-limitation benefits, the routing algorithm
ought to locate an ideal arrangement inside a sensible
I. INTRODUCTION
time. With a specific end goal to take care of this issue,
latest studies on such issues appear to concentrate on
A MANET (mobile ad-hoc network) might be
heuristic algorithms. Among heuristic algorithms, the
presented as a foundation less dynamic network which
comprehensive search algorithm is a general strategy
is an accumulation of free number of mobile nodes that
that comprises of deliberately specifying all
can convey to each other by means of radio wave. A
conceivable possibility for finding a worldwide answer
MANET is a self-arranging base, less networks of
for NP-difficult issues, and the routing optimization
mobile gadgets associated by remote. It is an
issue can likewise be comprehended by utilizing this
arrangement of remote gadgets called remote nodes,
algorithm. Be that as it may, finding the ideal
which progressively associate and exchange data.
arrangement by utilizing this algorithm requires
Every node in a MANET is allowed to move
inordinate computational time. To evade numerical
autonomously in any bearing, and will thusly change
troubles and lessen the computational weight,
its connections to different gadgets much of the time;
endeavors have been dedicated in discovering superb
each must forward activity inconsequential to its own
arrangements in a sensible computational time by
utilization, and along these lines be a switch. The
meta-heuristic optimization methods instead of finding
MANET network empowers servers and customers to
a worldwide arrangement [3]. Despite the fact that
convey in a non-altered topology zone and it's utilized
meta-heuristic algorithms don't promise accomplishing
as a part of an assortment of utilizations and quickly
the ideal arrangement, these algorithms will
developing networks [1]. With the expanding number
accomplish a worthy arrangement in a sensible time.
of mobile gadgets, giving the registering force and
Genetic algorithms (GAs) and Tabu search
network to run applications like multiplayer diversions
(TS) techniques are usually seen as in view of various
or cooperative work devices, MANETs are getting
establishments and points of view [2]. Albeit critical
increasingly vital as they meet the prerequisites of
contrasts between the techniques exist, the two
today's clients to associate and connect suddenly.
methodologies share certain components in like
Routing is a procedure of exchanging packets
manner that frequently have been neglected and left
from source node to destination node with least cost.
unexploited. The topic of this paper is to recognize
Consequently the algorithm for routing needs to sort
associations and differences between the strategies that
out and convey the data about the network states.
offer a ripe means for making crossover systems. The

334 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
International Journal of Computer Science and Information Security (IJCSIS),
Vol. 14, No. 10, October 2016

improvement of GA/TS crossovers offers a practically network is known in advance. The Tabu search [16, 17]
undiscovered region for exact research. The various is a scientific optimization strategy that has a place
effective uses of both GA and TS approaches with the nearby search strategies. It upgrades the
unequivocally contend for the value of researching execution of the nearby search technique by utilizing
brought together methodology. We propose ways that memory structures. The Tabu records contain traits
the varying systems of genetic algorithms and Tabu that are a great deal more successful, yet they raise
search can be favorably accommodated, welcoming another issue. Ordinarily, more than one arrangement
research and computational studies to recognize the is pronounced as Tabu and some of these
best approaches to coordinate these techniques. arrangements may have brilliant qualities that may
have not yet been gone by. To beat this issue, the
II. RELATED WORKS yearning criteria are acquainted with permit the
abrogating Tabu condition of the arrangement. The
In wired networks, traditional routing generally utilized yearning paradigm is permitting the
protocols are generally taking into account join state utilization of preferred arrangements over the best as
or separation vector algorithms, for example, the of now known arrangements. The genetic algorithm
Dijkstra algorithm or the Bellman-Ford algorithm. [18, 19] comprises of effective and broadly relevant
Vasilakos et al. [4] proposed a computational stochastic search and optimization strategies in view
knowledge approach for advancing the routing in of the standards of advancement hypothesis. This
various leveled ATM networks. This methodology algorithm has gotten impressive consideration
plans to productively apportion the network assets concerning its potential as an upgrading procedure for
while guaranteeing the nature of administration complex issues. It has been effectively connected to
necessities for every association. In remote networks, mechanical designing, for example, a fuzzy
perhaps mobile, multi-jump network, diverse framework, which can be alluded in neuro framework,
methodologies are required. Routing protocols ought bio savvy framework, and encompassing knowledge
to be dispersed, have low overhead, act naturally applications [20, 21]. Vasilakos et al. [22] proposed a
designing and have the capacity to adapt to habitually fuzzy set rationale based genetic algorithms for
changing network topologies [5]. Because of this between space routing of broadband network
prerequisite, an expansive number of routing protocols associations with nature of administration
have as of late been created for ad-hoc networks. prerequisites in a coordinated ATM and SDH
These protocols for mobile ad-hoc networks are networking engineering.
generally characterized [6] as either table-driven or
proactive protocols, which do attempt to keep precise III. PROPOSED WORK
data in their routing tables, or as on-interest protocols, A. Genetic Algorithm (GA)
which just build routing tables when information is GA is a developmental procedure for huge
sent to a destination. The destination sequenced space search. The general strategy of GA search is as
separation vector [7] and the remote routing protocol per the following:
[8] are well known case of table-driven protocols.
1) Population generation: A population is an
Dynamic source routing [9], on-interest separation arrangement of chromosomes and each speaks to a
vector routing and associativity-construct routing [10] conceivable arrangement, which is a mapping grouping
are illustrative in light of interest protocols. Vitality amongst undertakings and machines. The underlying
effectiveness is an imperative thought for mobile ad- population can be created by other heuristic
hoc networks. For vitality effectiveness, it might be algorithms, for example, Min-min (called seeding the
alluring to utilize not just a solitary path between a population with a Min-min chromosome).
source and a destination however to investigate
different paths. The traditional various routing plans 2) Chromosome assessment: Each chromosome is
connected with a fitness worth, which is the makespan
[11–13] have additionally been considered in mobile
of the assignment machine mapping this chromosome
ad-hoc networks. Deciding the improved path amongst
speaks to. The objective of GA search is to discover
source and destination nodes under a few compels is the chromosome with ideal fitness esteem.
an essential key in numerous routing protocols. In any
case, by utilizing traditional algorithms, it is hard to 3) Crossover and Mutation operation: Crossover
discover an answer inside an adequate calculation time operation chooses an arbitrary pair of chromosomes
on the grounds that if there are a substantial number of and picks an irregular point in the main chromosome.
nodes in the network, then the routing optimization For the segments of both chromosomes starting there
issue is NP-hard. to the end of every chromosome, crossover trades
Some routing protocols for deferral tolerant machine assignments between comparing
networks have likewise been proposed to overcome undertakings. Mutation haphazardly chooses a
chromosome, then arbitrarily chooses an errand inside
continuous, long-span availability disturbances. They
the chromosome, and haphazardly reassigns it to
are arranged into three sorts: deterministic, upheld, another machine.
and astute methodology [14, 15]. The deterministic
methodology can be composed when the data of

335 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
International Journal of Computer Science and Information Security (IJCSIS),
Vol. 14, No. 10, October 2016

4) Finally, the chromosomes from this changed nodes. The aggregate cost is the entirety of cost of
population are assessed once more. This finishes one individual bounces. The objective is to discover the
cycle of the GA. The GA stops when a predefined path with least aggregate cost between source node Vs
number of advancements are come to or all and destination Vd, where Vs and Vd have a place with
chromosomes unite to the same mapping V. This paper displays the productive on-interest,
This genetic algorithm arbitrarily chooses source started routing algorithm utilizing Tabu search
chromosomes. Crossover is the procedure of swapping algorithm and genetic algorithm. It is actualized as two
certain sub-successions in the chose chromosomes. modules viz., Network generation or introduction of
Mutation is the procedure of supplanting certain sub- routing tables utilizing Genetic algorithm and
successions with some undertaking mapping decisions generation of ideal path utilizing Tabu search
new to the present population. Both crossover and algorithm. At long last information is sent along the
mutation are done randomly. After crossover and created path.
mutation, another population is produced. At that point
this new population is assessed, and the procedure D. Hybrid TS/GA Algorithm
begins once again until some halting criteria are met.
The ceasing criteria can be, for instance, 1) no change Hybridization is a pattern in numerous chips
in late assessments; 2) all chromosomes meet to the away at meta-heuristics recent years. It exploits the
same mapping; 3) a cost bound is met.
advantages accumulated from various meta-heuristics,
B. Tabu Search (TS) to such an extent that the essential meta-heuristics are
simply canvas, beginning stages to start to take care of
TS is actualized starting with an irregular an optimization issue [23]. In the figures beneath, we
mapping as the underlying arrangement, produced propose an optimization technique in view of a half
from a uniform dissemination. To control the present and half approach in Fig.1.
arrangement and travel through the arrangement space,
Pseudo code for Hybrid Algorithm
a short bounce is performed. The natural reason for a
short jump is to discover the closest neighborhood least Algorithm 1:
arrangement inside the arrangement space. The 1. Central node generates n initial solutions using GA.
fundamental method to perform a short bounce is to It runs GA for fixed number of iterations, t.
consider, for every conceivable pair of undertakings, a. Select first population of fixed size and set j=1
every conceivable pair of machine assignments, while b. While(j<=t) Begin i. Apply the operator on the two
alternate assignments are unaltered. On the off chance parent schedules selected randomly to create two
that the new makespan is a change, the new offspring and replace the parents by the best two the
arrangement is spared, supplanting the present four schedules. ii. J=j+1 End
arrangement. The short bounce technique closes when 2. Central node leads m best solutions picked to the m
(1) each pair-wise remapping mix has been depleted remote worker nodes
with no change, or (2) the point of confinement on the 3. Each worker node runs the TS algorithm by means
aggregate number of fruitful jumps (limithops) is come of the first state received.
to. At the point when the short bounce method closes, 4. Upon receiving a converged result from one of the
the last mapping from the nearby arrangement space worker nodes, the central node stops execution
search is added to the Tabu rundown. The Tabu
rundown is a technique for monitoring the areas of the In GA, an initial population consisting of a set
arrangement space that have already been searched. of solution is chosen and then the solutions are
Next, another arbitrary mapping is created, and it must estimated. Comparatively more effective solutions are
vary from every mapping in the Tabu rundown by at nominated to have more off springs which are, in some
any rate half of the machine assignments (an effective way, related to the novel solutions. If the genetic
long jump). The instinctive motivation behind a long operator is selected correctly, the final population will
bounce is to move to another area of the arrangement have better solutions. GA expands the entire
space that has not already been searched. After each population. TS aim at producing one best solution. For
effective long bounce, the short jump method is the TS, we require several good initial solutions to
rehashed. The ceasing rule for the whole heuristic is confirm the necessary number of good early solution.
the point at which the entirety of the aggregate number The pseudo-code for the GA and GTA is given
of fruitful short jumps and effective long bounces Algorithm 1.
squares with limithops. At that point, the best mapping
from the Tabu rundown is the last reply.

C. Problem Definition

The network under thought is spoken to as G


= (V, E), an associated diagram with N nodes. The
metric of optimization is cost of path between the

336 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
International Journal of Computer Science and Information Security (IJCSIS),
Vol. 14, No. 10, October 2016

Performance Comparison - (Tabu with GA) Vs (GA) Vs (Simulated Annealing)


3500
Tabu with GA
Tabu
3000
Simulated annealing

2500

Minimum Routing Cost


2000

1500

1000

500

A B C D
Problem

Fig. 2(a): Minimum Routing cost

Fig.1: TS/GA Hybrid algorithm Performance Comparison - (Tabu with GA) Vs (GA) Vs (Simulated Annealing)
3.5

IV. RESULTS AND DISCUSSION 3


Tabu with GA
Tabu
Simulated annealing
In this area, we contrast the proposed Tabu
Average Execution Time(Sec)

2.5
search algorithm and two meta-heuristic algorithms,
the genetic algorithm and the simulated annealing, by 2

means of PC tests. All the examinations are executed


with the utilization of MATLAB, and they are 1.5

performed on a 1.80 GHz Pentium®4 that is furnished


1
with a Windows OS, 8 GB of memory and an Intel®
processor. The algorithms were connected to enhance 0.5
the routing issues with four distinctive network
topologies. The topologies are called issues A, B, C 0
A B C D
and D. Every issue contains a few nodes and Problem
connections, as is appeared in Table 1. Besides, the Fig 2(b): Average Execution Time
parameters of the algorithms that are utilized as a part
of the tests are appeared in Table 2. Fig. 2(a) shows the comparisons of the
Table 1: Problems for the experiments
minimum routing cost of three meta-heuristic
Problem Num. of Nodes Num. of links
algorithms: Tabu with genetic search, genetic
A 80 225 algorithm and simulated annealing. We applied three
B 160 456 algorithms to the given problems according to the
C 320 923 following conditions. In the Tabu with genetic search,
D 640 1862 Kmax is set to 100. In the genetic algorithm, p and q are
set to 1 and 0.25, respectively, and in the simulated
Table 2: The parameters of the meta-heuristic algorithms
annealing, β is set to 0.5. We see that in terms of the
Algorithms Param Values routing cost, the Tabu with genetic search algorithm
eters outperforms the genetic algorithm and the simulated
Tabu with Genetic K max 10/50/100/200 annealing. The Tabu with genetic search and the
algorithm Search simulated annealing algorithms activity the same
Genetic Algorithm Np 50 neighborhood generating operations, which are the
P 1/0.5 remove and the replace operations, but since the Tabu
Kmax 100 with genetic search regularly produces more next
Simulated Annealing Tinit 0.1 generation solutions than the simulated annealing, we
T final 0.00005 observe that the Tabu with genetic search has better
α 0.1 performance than the simulated annealing in the figure.
β 0.1/0.3/0.5/0.7/0.9 We also see that the Tabu with genetic search
Kmax 100 outperforms the genetic algorithm with the crossover

337 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
International Journal of Computer Science and Information Security (IJCSIS),
Vol. 14, No. 10, October 2016

and the mutation processes. This highest reason is that [3]. Zhao, W., Ammar, M., & Zegura, E. (2004).
the Tabu with genetic search with regularly generating A message ferrying approach for data delivery
neighborhood solutions is better useful to the routing in sparse mobile ad hoc networks. In
problem than the genetic algorithm with stochastically Proceedings of ACM/IEEE MOBIHOC
generating results. Fig. 2(b) shows evaluations of the [4]. Vasilakos, A., Saltouros, M., Atlasis, A., &
average execution time for the meta-heuristic Pedrycz, W. (2003). Optimizing QoS routing
algorithms in the given problems. The results show in hierarchical ATM networks using
that the Tabu with genetic search takes a shorter time computational intelligence techniques. IEEE-
than the genetic algorithm and the simulated annealing. Transactions on Systems, Man and
This is because the Tabu with genetic search reaches Cybernetics, Part C, 33(3), 297–312.
the optimal solution faster than other algorithms. [5]. Akkaya, K., & Younis, M. (2005). A survey
Moreover, though the Tabu with genetic search on routing protocols for wireless sensor
produces more neighborhood solutions for the current networks. Ad hoc networks, 3, 325–349.
solution than the simulated annealing, it takes a [6]. Royer, E. M., & Toh, C. K. (1999). A review
smaller time than the simulated annealing that of current routing protocols for ad-hoc mobile
randomly produces a next generation for the current wireless networks. IEEE Personal
solution. Communications, 6, 46–55.
[7]. Perkins, C. E., & Bhagwat, P. (1994). Highly
V. CONCLUSION dynamic destinationsequenced distance-vector
routing (DSDV) for mobile computer. In
In this paper, we proposed a routing Proceedings of the ACM SIGCOMM (pp.
optimization algorithm to productively decide an ideal 234–244).
path from a source to a destination in mobile ad-hoc [8]. Raju, J., & Garcia-Luna-Aceves, J. J. (2000).
networks. The proposed algorithm was planned by A comparison of ondemand and table driven
utilizing the tabu with genetic search algorithm, which routing for ad-hoc wireless networks. In
is an ordinary meta-heuristic algorithm. We initially Proceedings of ICC.
depicted the generally utilized operations, which are [9]. Maltz, D. (2001). On-demand routing in
encoding arrangement, the creation of the underlying multi-hop wireless ad hoc networks. PhD
arrangement and the repair capacity, and we thesis, Carnegie Mellon University, Pittsburgh,
introduced the area producing operations of the PA.
algorithm that incorporates the end condition. To look [10]. Perkins, C. E., & Royer, E. M. (1999). Ad-hoc
at the performance of the proposed algorithm, we on-demand distance vector routing. In
additionally built up the genetic algorithm and the Proceedings of the 2nd IEEE workshop on
simulated annealing that are the agent meta-heuristic mobile computing systems and applications
algorithms. We assessed the performance of the (pp. 90–100).
algorithms via doing the analyses by shifting the [11]. Toh, C. K. (1999). A novel distributed routing
quantity of nodes and connections, and we contrasted protocol to support ad hoc mobile computing.
the proposed algorithm and other meta-heuristic In Proceedings of IEE 15th annual
algorithms regarding the routing cost and the normal international conference on computers and
execution time for the routing issue. The examination communications (pp. 460–486).
results demonstrated that the tabu with genetic search [12]. Nasipuri, A., & Das, S. R. (1999). On-demand
beats different algorithms regarding the routing cost multipath routing for mobile ad hoc networks.
and normal execution time under different In Proceedings of the 8th ICCCN.
requirements, and it is appropriate for adapting the [13]. Valera, A., Seah, W. K., & Rao, S. V. (2003).
routing optimization issue. Cooperative packet caching and shortest
multipath routing in mobile ad hoc networks.
REFERENCES In Proceedings of IEEE INFOCOM.
[14]. Spyropoulos, T., Rais, R. N. B., Turletti, T.,
[1]. Jaspal Jindal and Vishal Gupta, Fuzzy Obraczka, K., & Vasilakos, A. (2010).
Improved Genetic Approach for Route Routing for disruption tolerant networks:
Optimization in MANET, International taxonomy and design. In Wireless networks
Journal of Advanced Research in Computer (pp. 1–22).
Science and Software Engineering, Volume 3, [15]. Pedrycz, W., & Vasilakos, A. (2001).
Issue 6, June 2013. Computational intelligence in
[2]. FRED GLOVER,~f JAMES P. KELLY~ and telecommunications networks. USA: CRC
MANUEL LAGUNA§, GENETIC Press
ALGORITHMS AND TABU SEARCH: [16]. Glover, F. (1989). Tabu search, Part I.
HYBRIDS FOR OPTIMIZATION, ORsimulated Annealing Journal on
Computers Ops Res. Vol. 22, No. 1, pp. 111- Computing, 1, 190–206.
134, 1995.

338 https://sites.google.com/site/ijcsis/
ISSN 1947-5500
International Journal of Computer Science and Information Security (IJCSIS),
Vol. 14, No. 10, October 2016

[17]. Glover, F. (1990). Tabu search, Part II. University and M.E degree in Computer Science and
ORsimulated Annealing Journal on Engineering from Rajas Engineering College under
Computing, 2, 4–32. Anna University in 2005 and 2009, respectively. He is
[18]. Holland, J. (1975). Adaptation in natural and currently pursuing his research in MAHER University.
artificial systems. Ann Arbor: Univ. of His research interests include Mobile Computing,
Michigan Press. MANET, Cloud Computing and so forth.
[19]. Goldberg, D. E. (1989). Genetic algorithms
in search, optimization & machine learning.
Reading: Addison–Wesley.
[20]. Pedrycz, W., & Vasilakos, A. V. (1999).
Linguistic models and linguistic modeling.
IEEE Transaction on Systems, Man, and
Cybernetics-Part B: Cybernetics, 29(6), 745–
757.
[21]. Acampora, G., Gaeta, M., Loia, V., & Dr S. Thabasu Kannan received his
Vasilakos, A. V. (2010). Interoperable and M.Tech degree in Computer Science and Engineering
adaptive fuzzy services for ambient from PRIST University. He received his Ph.D in the
intelligence applications. ACM Transactions area of Bioinformatics in Madurai Kamaraj
on Autonomous and Adaptive Systems, 5(2). University. He worked as HOD, Principal and
[22]. Vasiliakos, A., Ricudis, C., Anagnostakis, K., Director in different Engneering colleges. Currently he
Pedrycz, W., & Pitsillides, A. (1998). is working as Director in Theni College of Arts and
Evolutionary-fuzzy prediction for strategic Science. He published 49 research papers in
QoS routing in broadband networks. In International Journals. He published 28 papers in
Proceedings of IEEEFUZZ (pp. 1488–1493). confereences. His 5 research papers were published in
[23]. Bouzeboudja H., Chaker A., Allali A., research books. He got 8 prestigious awards for
Naama B., Economic Dispatch Solution academic and research excellence like Glory of
Using a Real-Coded Genetic Algorithm, Acta Education Excellence Award, Bharat Vidya
Electrotechnica et Informatica, 2005, 5(4), p. Shironmani Award, Eminent Educationist Award,
1-5. Golden Educationist of India Award, Rashtriya Vidya
[24]. Naama B., Bouzeboudja H., Ramdani Y., Saraswathi Puraskar award, Mother Teresa Sadbhavna
Chaker A., Hybrid Approach to the Award, Bharat Vidya Ratan Award, Life time
Economic Dispatch Problem Using a Genetic education achievement award from different
and a Quasi-Newton Algorithm, Acta organizations. He published 6 books in different field.
Electrotechnica et Informatica, 2008, 8(3), p. His research interests include Clustering and
31-35. modeling, pattern significance, privacy preserving,
watermarking database, Similarity search, diagnostics,
drug design. Health care cloud, Educational domain
and so forth.

Mr T.Prince received his B.E degree


in Electronics and Communication Engineering from
Government Engineering College under Anna

339 https://sites.google.com/site/ijcsis/
ISSN 1947-5500

You might also like