A CLUSTER BASED STABLE ROUTING PROTOCOL USING BINARY PARTICLE SWARM OPTIMIZATION TOWARDS SERVICE DISCOVERY IN MOBILE AD HOC NETWORK

International Journal of Mobile Network Communications & Telematics ( IJMNCT) Vol. 3, No.4, August 2013
DOI : 10.5121/ijmnct.2013.3404 29
A CLUSTER BASED STABLE ROUTING PROTOCOL
USING BINARY PARTICLE SWARM OPTIMIZATION
TOWARDS SERVICE DISCOVERY IN MOBILE AD HOC
NETWORK
M.Buvana1
and M.Suganthi2
1
Department of Computer Science and Engineering, PSNACET,Dindigul,Tamilnadu,
India
bhuvana_beula@yahoo.co.in
2
ECE,TCE,Madurai,India.
ABSTRACT
A Mobile Ad hoc Network consists of active nodes that can be in motion freely. These networks can be
improved by the use of clusters because of huge congestion in the whole network. In such a system, the
performance of MANET is improved by splitting the whole network into various clusters. The performance
of clustering is improved by the cluster head selection and number of clusters. In this paper, we have
designed a new protocol, Cluster based Stable Routing protocol (CSR) based on Binary Particle Swarm
Optimization (BPSO). The proposed algorithm shows how BPSO can be useful in enhancing the
performance of service discovery in MANET. This algorithm is to select the best cluster head and minimum
number of clusters. Experiments were performed to prove that the proposed method is an efficient method
for clustering in MANET. Also the path stability is combined with clustering with use of remaining battery
power of the nodes.
KEYWORDS
Mobile Ad hoc Network, stable routing protocol, Binary particle swarm optimization, clusters, cluster
head.
1. INTRODUCTION
Mobile ad hoc networks (MANETs) play a critical role in today’s wireless ad hoc network
research. A MANET consists of wireless nodes which can communicate with each other over a
wireless medium. Such a network can be formed spontaneously whenever devices are located
within transmission range. New nodes can join or leave an existing node dynamically, because of
the mobility in wireless networks. MANETs provide unforeseen applications in some new fields.
One such problem is how to create an organizational structure amongst these nodes . The
MANETs is the capability to organize them in an ad hoc method, as it is not feasible to arrange
these nodes into groups. For this reason, we need to group nodes into clusters[1],[2]. The
clustering phenomenon, as we shall see, plays an important role not just in organization
of the network but also can dramatically affect network performance. There are several key
limitations in MANETs that clustering schemes must consider, such as limited energy, network
lifetime, mobility and stability.
Different existing algorithms for the creation of clusters are limited in many routing protocols, as
well as clustering algorithm, such as (i) lowest-ID (ii) node-weight heuristic.

30
Lowest-ID [3] is known as identiﬁer based clustering. In this algorithm, ID is assigned to each
node. Periodically, every node broadcasts the list of nodes so that it can be heard. Among the
neighbors, the node with minimum ID is selected as a cluster head. The drawback of this
algorithm leads to battery drainage of certain nodes. Furthermore, it does not try to balance the
load equally across all the nodes.
Weighted clustering algorithms (WCA) [4] depend on the weight of the nodes in the network. In
these algorithms, the node with lower weight among the neighbors is selected as a cluster head.
Although it can increase the stability of clusters, it cannot obtain the optimal number of cluster
heads. Furthermore, calculation and storage of the weight are costly and the overhead induced by
WCA is very high.
In [5] this research, the main objective is towards maximizing the lifetime of wireless ad hoc
networks through energy efficient routing. Furthermore, the factors that control the reduction of
energy in ad hoc networks are shown and the method for load balancing, which is considered as
important factor in MANETs, is proposed. But as in many researches, the proposed method for
load balancing is the theoretical one and hence the correctness of the method is not guaranteed
since there are no results of simulation.
Li Q, Aslam J, Rus D [6] describes Online Power-aware Routing in Wireless Ad-hoc Networks.
This paper describes online power-aware routing in large wireless ad-hoc networks for
applications where the message order is not known. The main purpose of this research is to
optimize the lifetime of the network. Here an approximation algorithm called max-min z Pmin that
has a good empirical competitive ratio is proposed. To ensure scalability, a second online
algorithm for power-aware routing is introduced. This hierarchical algorithm is called zone-based
routing.
In this paper, we propose a routing protocol towards service discovery protocol for MANET,
cluster based stable routing protocol (CSR) based on binary particle swarm optimization (BPSO).
The main objective is to minimize the number of cluster, election of cluster head with efficient
energy consumption and maximizing the network lifetime. The rest of this paper is organized as
follows. Section 2 consists of power aware routing that includes the nodes which are having the
optimized power in the path. Section 3 introduces definitions and models like system model,
energy model and transmission range. Section 4 provides the proposed work with Novel Binary
Particle swarm optimization is applied to optimize the cluster head election procedure. The
simulation results have demonstrated its performance improvement in terms of packet delivery
ratio and network lifetime is explained in section 5.
2. POWER AWARE ROUTING
Commonly routing is the very important part in MANET. Until now, so many protocols have
been proposed for routing. The working of each protocol is different from each other and each has
their own benefits and limitations.The purpose behind this method is to send the message as
quickly as possible. But this should not be considered as the best path at all times. Because if the
battery power of any node in the selected path is very low then the message cannot be transmitted
through that path at any cost because in the middle of the transmission, the link may get broken
down due to insufficient power. In the network shown below, consider the path 1-6-3 as the
shortest path. In Fig 1, node1 starts transmission then the power gets depleted to some extent.

31
Fig. 1. Status of power during initial transmission
Fig. 2. Status of power after first transmission
Fig. 3. Status of power before final transmission
So this problem has to be resolved by the following ways
(a) The path has to reconstructed
(b)The message has to be retransmitted.
But here the power of the nodes got wasted unnecessarily. Now all these problems occur because
of the instability of the selected paths. Hence, the lasting battery power of the nodes in the
network should be taken into account during the route discovery process in MANETS to avoid
such link breakage during the message transmission. This is the concept of power aware
algorithm proposed so far.
3.SYSTEM MODEL AND DEFINITIONS
3.1.System Model
We consider a distributed heterogeneous wireless network where data tends to be correlated in
both time and space and nodes are assumed to be time synchronized. The nodes are organized
into clusters based on power calculation and distance based and spatial correlation. Each cluster

Interna
has an
topolog
networ
, each
nodes i
}An ex
cluster
3.2 En
All the
energy
during
where
When
selectio
receive
model:
Where
wavele
the tran
The re
detecte
correct
The tot
ational Journal o
aggregator o
gy is modele
rk and E repre
cluster head
in Ci are repre
xample of N
s are connect
nergy Mode
e nodes in th
y level (Eth) an
route discov
it is stored i
the residual
on by the d
ed signal (r(t
:
PT and PR
ength, d is the
nsmitting and
eceiver sensit
ed is, PRmin as
t reception of
tal amount of
of Mobile Netw
or cluster hea
ed as an undi
esents an edg
CHi has k sp
esented by Zk
(Ci) cluster i
ted.
Fig. 4.
el
he given area
nd each node
very. The resi
in the routing
energy is les
estination. O
t)) strength w
TGTPRP
are transmi
e distance betw
d receiving an
losPath
tivity, the mi
s from (1). Th
f signals. The
f energy consu
work Communic
ad denoted a
irected graph
ge which conn
patially corre
k = { Zk ;k=1
is shown in F
. Example of cl
a have same
must mainta
idual battery
g tables to m
s than thresh
On receiving
which holds t
2
t
h
RGT






d
r
h
itter power a
ween the send
ntenna respect
R
P-
T
Pss 
inimum recei
he receiver st
sender uses t
umed per tran
cations & Telem
s CHi and a
h G where G
nects two nod
elated neighbo
1…n}, i.e., N(
Figure 4. The
luster with its m
transmit pow
ain the value i
energy value
make routing d
hold energy v
the RREQ
the following
2




and receiver
ders and the r
tively. Hence
ived power n
trength is the
this PR as PRm
nsmitted pack
matics ( IJMNC
set of nodes
= (C, E). C
des within a c
oring nodes.
(Ci) = { Zk Є
e clusters are
member N(Ci)
wer and each
in their routin
e can be obta
decisions bas
value, that no
the intermed
g relationship
power resp
receiver and G
the node calc
necessary for
e only one par
min for further t
ket is written a
CT) Vol. 3, No.4
as Zk Є Ci.
represents c
cluster. Within
The k spatial
Ci |( Zk, CHi)
e formed suc
h node select
ng table to sel
ained to the n
sed on the ba
de is avoided
diate nodes c
p for two-ray
(1)
ectively, λ i
GT GR are the
culates the pa
(2)
r a signal to
rameter whic
transmissions
as
4, August 2013
32
The network
clusters in the
n a cluster C
lly associated
) Є E}U{ CH
ch that all the
ts a threshold
lect the nodes
network layer
attery energy
d in the route
calculates the
y propagation
is the carrier
e unity gain of
ath loss using
be correctly
ch decides the
s.
2
k
e
Ci
d
Hi
e
d
s
r
.
e
e
n
r
f
y
e

33
b
L/R*
T
P (3)
Where
α = transmitted energy
PT = transmitter power
L = packet length
Rb = data rate or bandwidth
The total amount of energy consumed per received packet is written as
b
L/R*
minR
P (4)
where, β, PRmin are the energy consumed in received packets and minimum receiving power
respectively.
The node then calculates the residual energy ER using the following parameters:
Σ – Initial energy taken by the node
α – Energy consumed in transmitting packets
β – Energy consumed in receiving packets
γ – Energy consumption in idle state.
  RE (5)
3.3.Transmission Range
Distance and the transmission power are generated randomly for the nodes. Average distance and
mean transmission power are calculated. Max distance of the nodes generated is taken as the
radius for coverage area in transmission range. Average transmission range is computed with
distance of the nodes.
(
minR
P
r
G
t
G
t
P
2
c )f4C 
tx
TR (6)
This transmission range is taken as the parameter to decide the cluster head. If a node has this
value with all other nodes, that node will be selected as the cluster head.
4.PROPOSED WORK
4.1.Problem Statement
In this proposed work, the remaining battery power of the node is calculated and then routing
decision will be made. And identify the nodes those are having the higher capacity. Based on the
power of the node cluster head (CH) is elected then the clusters are formed based on the
transmission range that is calculated as per equation (6). BPSO is applied to optimize the number
of clusters as well as manage the load among the clusters.

34
4.2.Method Overview
In this method, during route discovery, the remaining battery power of each node is first
calculated and then it is compared with a pre determined threshold value. A node is approved
only if it is greater than threshold value (Eth) to make sure that there is enough power for the node
to forward the packets to its neighbours. In fig 5, the power of node 6 is compared with the
threshold and it is low.
Fig. 5. Checking power condition for the first neighbor of source
So remove all the paths involving 6 and move to next neighbour which is node 2 as shown in fig
5
Fig. 6. Checking power condition for the next neighbor of source
Node 2 also fails, and node 5 satisfies the condition and now move to its neighbour node 4 as
shown in fig 6.
Fig. 7. Situation in which all the nodes of a path satisfies the power condition

35
If the condition fails, then that node can be simply removed from the consideration, not only the
node but also all the paths in which the node gets involved. So by this method, the entire route
discovery process is simplified by considering only one-third of the total number of nodes. In
addition to this, a special case has been dealt even though it is very rare. The following pseudo
code describes the steps about, when none of the nodes in the network meet the situation, then the
threshold value is decreased by half and the procedure is repeated until a node which satisfies the
condition is identified. As a result the stability of the paths in which the message has to be
transmitted gets increased and also it avoids unnecessary power reduction, which is one of the
very limited resources in MANETS.
/****** pseudo code for path stability *****/
Begin
BPi  initial power of nodes;
for each node ‘n’ in the network
calculate BPi;
L:
for i  0 to n
if(BP[Ni] >THi)
Forward to adja(Ni)
Else
Remove(Ni)
While(BP[Ni] != NULL)
THi = THi/2;
i++;
GoTo L;
Improve the path stability by avoid power depletion;
4.3.Novel Binary Particle Swarm Optimization
Kennedy and Eberhart proposed a discrete binary version of PSO for binary problems [3].In the
binary PSO, the particle's personal best and global best is modernized .The difference is that
velocities of the particles. This velocity must be controlled within the range of [0, 1]. The velocity
of each particle is obtained using the equation:
 
      
   
      

































11
1
,,,
,,22
,,11,
,
tvtxtx
txtgrc
txtprctv
tv
jijiji
jijibest
jijibestji
ji 
(7)
where vi,j is the velocity of particle, xi,j is the position of the particle, t is the number of iterations,
c1, and c2 are two positive constants, referred to as the cognitive and social acceleration factors
respectively, r1 and r2 are random numbers within the range [0,1], and w is the inertia weight.
The particle’s best position (pbest) is denoted as pibeset,j,the best position among all particles in the
swarm is denoted as gibeset,j.

36
   




























minmin
max
maxmax
)1(,
)1(,1
)1(,
1(
vtvifv
vtviftv
vtvifv
tvh
i
ii
i
i  (8)
Here vi depends on the sigmoid function
ivivsig 


1
1
)( (9)
Then the position of the particle is updated as[9]






)1(()(3,1
)1(()(3,0
)1(
tvsigtrif
tvsigtrif
tx
i
i
i (10)







1,
0,
,
0
,
,
1
,
,
jiji
jiji
ji
xifv
xifv
v


(11)
Where r3 is a random number within the range [0,1].
22
0
,22
0
,
22
1
,22
1
,
11
0
,11
0
,
11
1
,11
1
,
2,2,1
2,2,1
1,1,0
1,1,1
rcdandrcdThenpif
rcdandrcdThenpif
rcdandrcdThenpif
rcdandrcdThenpif
jiji
j
gbest
jiji
j
gbest
jiji
j
ibest
jiji
j
ibest




(12)
2,1,
1
,
1
,
1
,
1
, jijijiji ddwvv  (13)
2,1,
0
,
0
,
0
,
0
, jijijiji ddwvv  (14)
4.4.Cluster Formation
The proposed algorithm is a fully distributed algorithm as in section 3.1. All nodes in the mobile
network share the same responsibility and act as a cluster head. The proposed protocol takes
rounds where each round begins with a setup phase at the time cluster is formed. At the starting
phases, all the nodes share their energy level as per the energy model in section 3.2, to its
neighbor node. To ensure that only node with sufficient energy and transmission range are
selected as a cluster head (CH) as per section 3.3. The node with an energy above the threshold
value is suitable to be a cluster head for this round. During second phase, Novel Binary particle
swarm optimization is used to improve the performance of the cluster head election procedure.
The aim is that the numbers of clusters are minimized while the load in the network is balanced
among the clusters.

37
4.4.1.Cluster head election procedure.
For each particle in the population every dimension of the particle is traversed according to the
order of their experience. Cluster head election procedure is not a periodic and it is invoked at
starting and rarely when the network node joins in the cluster or moves away from the cluster
4.4.2.Initialize cluster head.
Cluster head candidate have the possible to become the cluster head for the cluster in the current
round. Cluster head candidate is selected based on the threshold energy (Eth) and based on the
transmission range as per equation (8).
Nn
C
pnCjZ pnCHjZ pndf
..2,1
max1










(15)
where zj member of the node in the cluster
CHp cluster head of the particle
||
pn
C Number of nodes that belongs to cluster cn of particle p.










2
11
))()(( jpn
s
jpn
pnCHjZ
zCHzCHd 




 N
n
pn
N
i
CHE
zE
f
i
1
1
)(
)(
2
(16)







tx
TRvudifrangeofout
tx
TRvudif
f
),(
),(r2
3

(17)
Where,
TRtx-Transmission Range based on the distance.
In this approach, a local lbest is found for each swarm member selected from the ‘closest’ two
swarm members. The distance between each particle and its neighbour’s particle that computed
the proximity depends on the first function, f1. Consequently, the preference of local optima
among the neighbours of a particle is based upon the second function, f2. The Third function f3
describes the optimal function for the transmission range through the nodes distance. In the next
sub-section, the steps of the algorithm will be described in detail.
/*Algorithm for cluster formation*/
1. Initialize swarm Xi, the position of the particles are initialized. Elements of Xi are
randomly selected from the binary values o and 1.
2. Initialize the randomly selected zk among the eligible cluster head candidate.
3. Particles position using the equation(7)
4. for each Xi in swarm find the function,
5. Assign each node ni, i =1,2……N in the network to the closest cluster head CH.
a. Assign each node ni,i=1,2……N in the network to the Closest cluster head CH.

38
b. Calculate the function f1, f2, f3 using (15)(16)(17).
6. Evaluate the performance of each particle to its best presentation.
if(f1(Xi(t+1)<f1(Pibest )and(f2(Xi(t+1)<f2(Pibest )and f3
f1(Pibest )= f1(Xi(t+1))
f2(Pibest )= f2(Xi(t+1))
f3(Pibest)= f3(Xi(t+1))
Pibest = Xi(t+1)
7. Compare the performance of each particle to the Gbest particle.
if(f1(Xi(t+1)<f1(Gibest )and (f2(Xi(t+1)<f2(Gibest )
f1(Gibest )= f1(Xi(t+1))
Gibest = Xi (t+1)
8. Apply the velocity update on all particles 10
j
vandjv

9. Apply the binary position update using (10)(11)
10. Calculate the velocity change of the bits j
cv

.
11. Generate the random variable rij in the range (0,1).Move each particle to new position
using (12).
12. Reinitialize M.
13. Repeat steps 2 to 11 until the maximum iteration is reached.
5.SIMULATION AND RESULTS
In this section, we evaluate the performance of our proposed technique via simulations. We The
network simulation model was built using MATLAB. The BPSO starts with a “swarm” of nodes
randomly generated. As shown in Figure 8 is a randomly deployed network with equal initial
energy of 0.25 J. A linear decreasing inertia weight value w is 0.8 and coefficients c1 and c2 both
are set to 2 as proposed in [7]. r1 and r2 are random numbers within the range [0,1]. In the
simulation wireless network composed by 400 nodes, for optimization, we have used 20 particles,
which are denoted by all nodes coordinates, for our experiment nodes are randomly distributed in
a 300m x300 m area square mobile ad hoc network shown in figure(8), and the maximum number
of iterations we are running is 6000. The transmission range of each node in the network and the
maximum speed (velocity) of the particle is set as 5 units and 50.Initially cluster head candidate is
selected based on the Residual energy (ER) which is calculated from the Eq. (5) and based on the
transmission range as per equation (7).Based on the transmission range nodes are moved and
located in their final position. For optimizing the clusters functions are calculated.
The nodes are organized into clusters by the cluster head candidate. Each particle will be
estimated by the functions in equations (15) (16) (17) in each iterations. Our purposed work is to
find the optimal location of cluster heads. As can be seen in Figure 9, our proposed algorithm
CSR-BPSO locates minimum number of clusters (3 clusters) in the whole network than
SEP[8],LEACH[9], PSO-C in the 300 x 300 m area with a transmission range of 35. The
proposed algorithm CSR-BPSO mechanism better than the other algorithms in terms of produce
the average number of clusters. The reason to find the best solution is that the total fitness value
should be minimal. The proposed work results confirm the network with minimum number of
clusters that can decrease the routing cost of the network.

Interna
For the
numbe
time. I
the pow
conside
is repe
useful
environ
ational Journal o
e current pap
er of clusters
nitially the re
wer of each
ered as failed
eated. Then th
to reduce t
nment. The
of Mobile Netw
per, Life time
are optimize
equired powe
node with th
d paths. If no
he message i
the number
clusters are
work Communic
Fig. 8. N
Fig. 9.optimiz
e of a node i
ed. Our analy
er is set as th
he threshold.
paths availab
is transmitted
of hops and
huge when
cations & Telem
Node Deployme
zed cluster form
is maximized
ysis focuses t
e threshold a
If it fails, all
ble, the thresh
d via the disc
d the delays
the transmis
matics ( IJMNC
ent
mation
d and the num
the energy ef
and the route
l the paths in
hold is reduce
covered path
s of packets
ssion ranges
CT) Vol. 3, No.4
mber of clust
fficiency and
is discovered
nvolving that
ed by half an
with enough
transmitted
of nodes ar
4, August 2013
39
ter heads and
d network life
d by checking
node will be
nd the process
h power. It is
in a cluster
re little. The
9
d
e
g
e
s
s
r
e

40
evaluation results show that the proposed method performs better than other algorithms in a
mobile ad hoc network environment.
The simulation results have been compared with various existing protocols. Table I describes the
Life Time of the node for different amount of energy and it also shows the comparisons of SEP,
LEACH, PSO-C with our proposed protocol CSR-BPSO algorithm. It shows when the first node
lost their energy during different energy levels 0.25J, 0.5J, 1J. Our proposed CSR-BPSO
algorithm has the highest lifetime compare to the other protocols.
Table 1 Lifetime of the Node
Protocol First Node Dies Last Node Dies Energy(J)
SEP 590 1654 0.25
LEACH 420 806
PSO-C 801 1809
CSR-BPSO 1011 2010
SEP 1001 2715 0.5
LEACH 934 2951
PSO-C 1501 3001
CSR-BPSO 1901 3125
SEP 1983 4100 1
LEACH 1851 3961
PSO-C 2002 4510
CSR-BPSO 2251 5805
The performance of the CSR-BPSO algorithm is compared with the well known algorithm
LEACH, SEP and PSO-C. In every round a set of new cluster heads is selected and the member
nodes send packet to their connected cluster head.
Figure 10 demonstrates the lifetime of the nodes in the clusters. The Number of nodes that stay on
active after each round. Initial energy is assigned as 0.25J.This figure shows that the network life
time for CSR-BPSO increases considerably compared to PSO-C, LEACH, and SEP. This
development is based on the following reasons, Firstly considerable energy saving is achieved by
CSR-BPSO algorithm through the use of dynamic clustering with the proposed functions. By
using the proposed dynamic clustering method, the CSR-BPSO algorithm can attain the minimum
number of clusters all the way through the simulations which leads to minimum energy.
Secondly, the functions can generate improved network partitioning with smallest amount
network energy dissipation. Here the total energy efficiency is raised nearly 48% than the PSO-C,
LEACH, SEP. After 5805 rounds only the proposed algorithm nodes are in dead situation, but in
the case of PSO-C, LEACH, SEP all the nodes are lost their energy nearly 806 rounds itself.

41
Fig. 10. Lifetime of the nodes
Figure 11 gives the results of amount of packet delivered in the network. Equation (18) shows the
ratio between number of packets received and number of packets transmitted. This parameter
finds the fraction of successful packet reception. In figure 11, the ratio of packet delivery in the
cluster for various numbers of nodes in Ad hoc network environment. In our algorithm, the ratio
of packet delivery is high by maintaining a stable route by considering residual energy and
transmission power taken from PHY and MAC layers. It reduces the rate of packet loss due to
link break by choosing an alternate path according to residual energy and transmission power.
Fig. 11. Packet Delivery Ratio vs Number of Nodes

42
Packet delivery ratio (18)
Furthermore, we also observed in the above Figures that the network lifetime increased to a
certain length in the cluster formation scenario. In our proposed work network life time will be
increased when compare to the LEACH, SEP, PSO-C. With this increase, the Network lifetime
was further prolonged when compare to the other cluster formation technique. This scheme can
produce results close to the optimal result.
6.CONCLUSION
In this paper, we have proposed a CSR-BPSO algorithm for routing towards service discovery
using a binary particle swarm optimization algorithm. Our algorithm improves the routing
function by taking into account of number of clusters and load among the clusters. It tries to
minimize the number of clusters as well as cluster heads.
The Performance results show a minimum number of clusters in every round during simulation.
Furthermore, the proposed algorithm gives the network lifetime is 84.34% and packet delivery
ratio is more compared to other algorithms such as LEACH, SEP, and PSO-C.
REFERENCES
[1]. Dhurandher.S.K., & Singh.G.V. (2006). “Power Aware Clustering Technique in Wireless Ad Hoc
Networks”, International Symposium on Ad Hoc and Ubiquitous Computing, ISAUHC '06.
[2]. Bandyopadhyay S, Coyle EJ. “An energy efficient hierarchical clustering algorithm for wireless
sensor networks”.In IEEE INFOCOM 2003; 1713–1723.
[3]. Kennedy, J.; Eberhart, R.C. (1997),” A discrete binary version of the particle swarm algorithm”, IEEE
Conference on Systems, Man, and Cybernetics, 1997.
[4]. Vieu .V.B., Nasser .N., & Mikou .N. (2006). “A Weighted Clustering Algorithm Using Local Cluster-
heads Election for QoS in MANETs.” IEEE GLOBECOM.
[5]. Natalia Vassileva, Francisco Barcelo-Arroyo, “A Survey of Routing Protocols for Maximizing the
Lifetime of Ad Hoc Wireless Networks”, International Journal of Software Engineering and Its
Applications, Vol. 2, No. 3, July, 2008
[6]. Nen-Chung Wang, Jong-Shin Chen, Yung-Fa Huang, Yu-Li Su, “A Power-Aware Multicast Routing
Protocol for Mobile Ad-hoc Networks With Mobility Prediction”, Wireless Pers Communication,
2007
[7]. Mojtaba Ahmadieh Khanesar, Hassan Tavakoli, Mohammad Teshnehlab andMahdi Aliyari
Shoorehdeli K N. Toosi” A Novel Binary particle swarm algorithm”, Mediterranean Conference on
Control & Automation, pp 1-6,.2007.
[8]. Georgios Smaragdakis, Ibrahim Matta & Azer Bestavros, (2004) “SEP: A Stable Election Protocol for
Clustered heterogeneous wireless Sensor Networks”, Second International Workshop on Sensor and
Actuator Network Protocols and Applications (SANPA -04).
[9]. Fan Xiangning & Song Yulin, (2007) “Improvement on LEACH Protocol of Wireless Sen-sor
Network”, Pro. International Conference on Sensor Technologies and Applications,EEE2007.

43
M.Buvana has over 9 years of teaching experience and is currently working as a Associate
Prof in PSNA College of Engineering and Technology. She finished her bachelor of
Engineering (computer science) at PSNA College of engineering and Technology, Dindigul
in the year 2004. And She completed her M.E in PSNA College of engineering and
Technology in 2008. She published two papers in International journal and she has
published papers in the 7 national and 5 International and IEEE conferences such as ICoAC
2011,ICICCA 2013,etc.. She is pursuing PhD degree in the field of Mobile communication at Anna
University Chennai, India.
M.Suganthi is currently working as Professor of Department of Electronics and
Communication Engineering in Thiyagaraja College of Engineering, Madurai. She
graduated her B.E. in Electronics and Communication Engineering from Thiyagaraja
College of Engireeing, Madurai, M.E. in Communication Systems from P.S.G College of
Technology, Coimbatore, and Ph.D in Mobile Com munication from Madurai Kamaraj
University, Madurai. She has been an educator in the technical teaching society for the past
23 years. She has published various papers about the wireless communication in different conferences and
journal and she has guided 6 research scholars.

A CLUSTER BASED STABLE ROUTING PROTOCOL USING BINARY PARTICLE SWARM OPTIMIZATION TOWARDS SERVICE DISCOVERY IN MOBILE AD HOC NETWORK

Recommended

More Related Content

What's hot (18)

Viewers also liked (9)

Similar to A CLUSTER BASED STABLE ROUTING PROTOCOL USING BINARY PARTICLE SWARM OPTIMIZATION TOWARDS SERVICE DISCOVERY IN MOBILE AD HOC NETWORK (20)

Recently uploaded (20)

A CLUSTER BASED STABLE ROUTING PROTOCOL USING BINARY PARTICLE SWARM OPTIMIZATION TOWARDS SERVICE DISCOVERY IN MOBILE AD HOC NETWORK