Int. J. Communications, Network and System Sciences, 2011, 4, 249-255
doi:10.4236/ijcns.2011.44030 Published Online April 2011 (http://www.SciRP.org/journal/ijcns)
Energy Efficient Computation of Data Fusion in Wire less
Sensor Networks Using Cuckoo Based
Particle Approach (CBPA)
Manian Dhivya1, Murugesan Sundarambal2, Loganathan Nithissh Anand3
1Department of EEE, Anna University of Technology Coimbatore, Coimbatore, India
2Department of EEE, Coimbatore Institute of Technology, Coimbatore, India
3Department of Mechanical Engineering, Anna University of Technology Coimbatore, Coimbatore, India
Received February 27, 2011; revised March 26, 2011; accepted March 30, 2011
Abstract
Energy efficient communication is a plenary issue in Wireless Sensor Networks (WSNs). Contemporary en-
ergy efficient optimization schemes are focused on reducing power consumption in various aspects of hard-
ware design, data processing, network protocols and operating system. In this paper, optimization of network
is formulated by Cuckoo Based Particle Approach (CBPA). Nodes are deployed randomly and organized as
static clusters by Cuckoo Search (CS). After the cluster heads are selected, the information is collected, ag-
gregated and forwarded to the base station using generalized particle approach algorithm. The Generalized
Particle Model Algorithm (GPMA) transforms the network energy consumption problem into dynamics and
kinematics of numerous particles in a force-field. The proposed approach can significantly lengthen the net-
work lifetime when compared to traditional methods.
Keywords: Cuckoo Search, Generalized Particle Model, Energy Efficiency, Clustering, Sensor Networks
1. Introduction
Wireless sensor networks (WSNs) facilitate innovative
applications and necessitate non-conventional paradigms
for protocol design. Sensor node is a tiny device that
includes three basic components: a sensing subsystem for
data acquisition from the physical surrounding environ-
ment, a processing subsystem for local data processing
and storage, and a wireless communication subsystem
for data transmission [1]. Their structure and characteris-
tics depend on their electronic, mechanical and commu-
nication limitations but also on application-specific re-
quirements [2].
Based on the architecture and power breakdown, sev-
eral approaches are considered simultaneously to reduce
power consumption in wireless sensor networks. The
main enabling techniques at a very general level are
Clustering Schemes, Routing Protocols, Duty Cycling
Schemes, Data Driven approaches and Mobility [3]. Clu-
stering, as potentially the most energy efficient organiza-
tion, has witnessed wide application in the past few years
[4] and numerous clustering algorithms have been pro-
posed for energy saving. Hence efficient data clustering
techniques must be used to reduce the data redundancy
and in turn reduce overhead on communication [5]. Low
Energy Adaptive Clustering Hierarchy (LEACH) and
Hybrid Energy Efficient Distributed clustering (HEED)
are the traditional approaches to effective data-gathering
protocols in WSNs. Energy consumption in multi-cluster
sensor networks is also explained [6]. A protocol pro-
posed in [7] optimizes the network performance under
the metric of information rate per Joule while ensuring a
given quality of service (QoS). QoS effectively exem-
plify the time and energy required to collect the pack-
ets .The traditional approach of repeated clustering
makes communication and processing overheads. Nu-
merous strategies have been projected so far, incorporat-
ing optimization techniques, but the need for energy
consciousness remain as a tough issue.
This paper proposes a new optimization based data
collection scheme incorporating Cuckoo Search (CS) and
Generalized Particle Model Algorithm (GPMA). The
formulated technique of combining the above algorithms
is stated as Cuckoo Based Particle Approach (CBPA).

M. DHIVYA ET AL. 250
The aspect of formulating this technique is incorporation
of constraints that helps in energy efficient data collec-
tion or fusion, by eliminating data redundancy and
minimization of energy consumption. Cuckoo Search is
performed to form sub-optimal data fusion chains. The
collected information is transmitted to the base station
via cluster heads through GPM algorithm. In the GPM
algorithm energy constraints are incorporated and the
information is routed in shortest path. The proposed
technique shows better performance in optimum number
of clusters, total energy consumption and prolonging of
network lifetime.
The rest of the paper is organized as follows. The
problem formulation and Energy Model is explained in
Section 2. Basics of Cuckoo Search and Generalized
Particle Approach Algorithm are explained in Section 3.
Cuckoo Based Particle Approach (CBPA), chain for-
mation, proposed methodology is explained in Section 4.
Results and analysis are described in Section 5. Conclu-
sions are drawn in Section 6.
2. Problem Formulation and Energy Model
The aim of this paper is minimization and conservation
of energy of WSNs. To minimize the usage of energy
sensor nodes are formed as static clusters in first phase,
by Cuckoo Search. The radio model adopted is, stated by
Heinzelman et al [8]. The following are the most widely
used assumptions and model in sensor network simula-
tion and analysis [9].
Nodes are dispersed randomly following a uniform
distribution in a 2-dimensional space and the location of
the Base Station (BS) is known to all sensors. The nodes
are capable of transmitting at variable power levels de-
pending on the distance to the receiver. The nodes are
unaware of their location. The nodes can estimate the
approximate distance by the received signal strength if
the transmit power level is known, and the communica-
tion between nodes is not subject to multi-path fading. A
network operation model similar to that of LEACH and
HEED [10] is adopted here, which consists of rounds.
Each round consists of a clustering phase followed by a
data collection phase.
2
electrical crossover
4
electrical crossover
. for 0
. for
fs
Tx
mp
lEdd d
E
lEdd d


 

 


(1)
The amount of energy consumed for transmission ETx,
of l-bit message over a distance d is given by Equation
(1).
electrical
.
Rx
ElE (2)
where Eelectrical is the amount of energy consumed in elec-
tronic circuits, εfs is the energy consumed in an ampli-
fier when transmitting at a distance shorter than dcrossover,
and εmp is the energy consumed in an amplifier when
transmitting at a distance greater than dcrossover. The en-
ergy expended in receiving a l-bit message is given by
the above Equation (2).
3. Background Work for Data Fusion
Cuckoo search (CS) is an optimization algorithm deve-
loped by Xin-She Yang and Suash Deb in 2009. It is a
novel algorithm which is inspired by the obligate brood
parasitism of some cuckoo species by laying their eggs in
the nests of other host birds of other species. It is more
efficient than Genetic Algorithm and Particle Swarm
Optimization to adapt to wider class of optimization pro-
blems [11].
In Figure 1 the Pseudo code for Levy Flights is given.
The quality or fitness of a solution is proportional to the
value of the objective function for the optimization
problem.
The CS is based on three idealized rules:
1) Each cuckoo lays one egg at a time, and dumps its
egg in a randomly chosen nest
2) The best nests with high quality of eggs will carry
over to the next generation
3) The number of available host’s nests is fixed, and
the egg laid by a cuckoo is discovered by the host
bird with a probability Pα. The worst nests are
discovered and dumped from further calculations.
X is related to the new solution, for an ith cuckoo and
then levy is performed as given in Equations (3) and (4).
 

1*Levy
ii
tt
X
X 
 
n (3)
 
1*
ii
tt
t
X
X 
E (4)
3.1. Generalized Particle Model Algorithm
Generalized Particle Model given by Dianxun Shuai [12],
finds its application in many fields of network oriented
applications. It is similar to Swarm Intelligence (SI)
technique. This algorithm eradicates the unknown em-
pirical performance and much computation time. For a
multi-objective optimization problem, the algorithm con-
trols the parameters in parallel or dispersedly.
The GP model consists of numerous particles and
forces, with each particle having its own dynamic equa-
tions to represent the network entities and force having
its own time-varying properties to represent various
social interactions among the network entities. Each par-
ticle in GP has four main characteristics [13]:
1) Each particle in GP has an autonomous self- driv-
ing force, to embody its own autonomy and the
personality of network entity.

M. DHIVYA ET AL.251
begin
Objective function f(x), x = (x1,..., xd)T
Generate initial population of
n host nests xi (i = 1, 2,..., n)
while (t < MaxGeneration) or
(stop criterion)
Get a cuckoo randomly by Levy
flights
evaluate its quality/fitness Fi
Choose a nest among n (say, j)
randomly
if (Fi > Fj),
replace j by the new
solution;
end
Fractions (pa) of worse nests
are abandoned and new ones
are built;
Keep the best solutions (or nests
with quality solutions);
Rank the solutions and find the
current best
end while
Post process results and visualization
End
Figure 1. Pseudo code for cuckoo search via levy flights.
2) The dynamic state of every particle in GP is a
piecewise linear function of its stimulus, to guar-
antee a stable equilibrium state.
3) The stimulus of a particle in GP is related to its
own objective, utility and intention, and to realize
the multiple objective optimizations.
4) There is variety of interactive forces among parti-
cles, including unilateral forces, to embody various
social interactions in networks.
The GPM is suitable for MAS’s problem-solving in
these more complex environment: multi-autonomy agents,
multi-type coordination, multi-objective optimization,
higher-degree parallelism, and random and emergent
events [14].The assignment matrix of task allocation and
resource assignment in MAS,



,
ik nm
StS t 

(5)
where, .
  
,,ζ
ijijij ij
St atptt
3.1.1. GP Algo r i thm
In Figure 2, the algorithm for Generalized Particle Mod-
el is given. The figure illustrates the necessary steps in
devising the model. A particle may be driven by numer-
ous forces that are produced by the force-field, other
particles and by own. The approach incorporates hybrid
energy functions to maintain the advantages of the tradi-
tional approaches and to eliminate the deficiencies of the
same. The gravitational force produced by the force field
tries to drive a particle to move towards boundaries of
the force-field, which embodies the tendency that a par-
ticle pursues maximizing the aggregate benefit of sys-
tems [15]. In Figure 3, the pushing or pulling forces
produced by other particles are used to embody social
coordination’s among agents. The self-driving force pro-
duced by a particle itself represents autonomy and per-
sonality of individual agents. Under the exertion of re-
sultant forces, all the particles may move concurrently in
a force-field [16]. In this way, the GPM transforms the
1. Initialization: (in parallel)
At time t = 0,




1, ,,1, ,.inkm
Agents A
ik
(t), payment P
ik
(t), intention strength


ik t

vertical coordinate of the particle q
ik
(t)
2. Calculation:
utility U
ik
(t), potential energy functions p(t), q(t), ψ
i
(t), fo
r
each partical s
ik
(t)
3. Condition 1:
If


dd0
ik
ut t

Hold for every particle, finish with success,
Else goto step 4
4. Condition 2:
Calcu late


dd0
ik
ut t

and update


ik
ut
Calculate


dd
ik
at t ,




1d d
ik ikik
A
tat att
Calculate


dd
ik
p
tt




1d d
ik ikik
Ptptp tt
Goto step 2.
Figure 2. Algorithm for generalized particle model.
Figure 3. Generalized particle model.

M. DHIVYA ET AL. 252
problem-solving process in Multi agent system (MAS)
into kinematics and dynamics of particles in the for-
ceield.When all the particles reach their equilibrium
states, the solution to the optimization of task allocation
and resource assignment is obtained.
4. Cuckoo Based Particle Approach (CBPA)
The energy function for the Cuckoo Search is designed
as [17];
 
1
1
100*
n
i
f
dfi di


 (6)
The above equation is derived from Equation (2), by
considering the value of εmp as 100 pJ/bit/m2 for n = 2,
i.e.; communication range between the sensors. The fit-
ness function F, is considered in this problem for mini-
mization of energy and maximization of lifetime of the
nodes.
4.1. Proposed Algorithm for Data Fusion
Input: A set of N sensor nodes in randomly deployed
field and a base station.
Step 1: Initialization:
Select the number of sensor nodes, cuckoo nests, eggs
in nests to start the search. Initialize the location and en-
ergy of nodes and the location of base station.
Step 2: Formation of Static Clusters:
The clusters are formed, by Cuckoo Search technique.
Each egg in a nest corresponds to a sensor node. A group
of M nests are chosen with N eggs in it. The probability
of choosing the best egg or quality egg is done by random
walk. Step size and Levy angle is updated in each itera-
tion. In turn the nests are updated. The optimal solution
i.e.; best egg – high energy node is taken as cluster head
in context to energy, distance between the nodes and dis-
tance to the base station.
The worse nets are abandoned in normal Cuckoo
Search. In order to compensate the unequal energy dissi-
pation, the worse nets (or) least energy nodes are allowed
to join the cluster as non cluster head nodes, in the pro-
posed approach. The less energy nodes join the proximity
cluster heads to form cluster. The cluster formation is
done by appropriate advertisement of cluster-head to all
other nodes to join a particular cluster. The cluster head is
not permanent. In each run, according to the residual en-
ergy of the nodes, the cluster head is periodically changed.
This helps to eradicate the communication overhead and
redundancy.
Step 3: Shortest Path Routing:
After the clusters are formed, the Cluster Heads (CHs)
fuse or aggregate the information before forwarding it to
the base station. The energy model incorporates free
space radio model followed by all nodes. The inter cluster
and intra cluster routing via shortest path is to be per-
formed based on the application. Intra cluster refers to
communication between cluster-head and non cluster-
head nodes within the cluster. Inter cluster communica-
tion refers to communication between the clusters. For
intra cluster communication, the most widely used meth-
odology as followed by the basic LEACH algorithm con-
cept is TDMA Scheduling- Time Division Multiple Ac-
cess Scheduling is followed.
For inter cluster communication, Generalized Particle
Model is used. The objective of using GPM approach is
optimization of route and extension of the network life-
time. GPM is given in many types according to the net-
work optimization problems by Dianxun Shuai. Normally,
communication between two Wireless Sensor Nodes
happens, when there is no other interfering node between
two nodes. It is assumed that there exists wireless path
and link between two nodes during communication.
The Generalized Particle Model transforms the net-
work shortest path problem into kinematics and dynamics
of numerous particles in a force-field. Nodes are consid-
ered as particles, and utility, overall utility, potential en-
ergy due to gravitational force, potential energy due to
interactive force of particles are calculated in each itera-
tion. It is personified in the described model, as the resul-
tant forces on the particles are high, the particles also
move fast. Then the particles are tested for stability con-
dition. If the particles are stable, then the algorithm is
terminated successfully. Else the particles are updated,
with the hybrid energy equations to obtain the optimal
solution.
The shortest path is calculated by the link cost of each
links. The link cost is substituted as the residual energy of
nodes; with context to the distance to the communicating
nodes. In more brief terms the residual energy of cluster
head to communicating cluster head is to be considered
for communication of the sensed data. After several
numbers of iterations, the optimal path to transmit the
data to the base station is being established.
The important steps in Cuckoo Based Particle Ap-
proach (CBPA) are listed in Figure 4.
5. Results and Analysis
In this section the performance of the proposed technique
is evaluated via simulation results. The network model is
simulated using MATLAB. The results are summarized
after running several iterations.
The focus on this paper, as by the objective function, is
minimization of energy and maximization of lifetime.
The simulation Parameters are listed in Table 1.

M. DHIVYA ET AL.253
Start
Initialization of parameters
Formation of Cluster Heads using Cuckoo Search
Calculate the distance of nodes to the base station
Formation of clusters by joining of Non
Cluster-Head nodes
Intra cluster communication by TDMA Scheduling
Inter cluster communication by Generalized
Particle Model
Stop
Figure 4. Flowchart of cuckoo based particle approach.
Table 1. Simulation parameters.
Sensor deployment area
Base station location
100 m *100 m
(50 m, 150 m)
Number of nodes 100 - 200
Data Packet size 100 bytes
Control Packet size 25 bytes
Initial Energy 2 J
Eelectrical 50 nJ/bit
εfs 10 pJ/bit/m2
εmp 0.0013 pJ/bit/m4
Mode of Topology configuration random
MAC Protocol IEEE 802.15.4
Duty cycle duration 1 second
Cuckoo step size 1
Round duration 60 seconds
In Figure 5, the clustering energy for hundred num-
bers of nodes is explained. It is seen the clustering energy
is less than 0.05 joules for LEACH, less than 0.04 joules
for HEED and less than 0.02 joules for CBPA. In Figure
6 and 7, the network lifetime of nodes is explained in
context to the first node death and rounds until the last
node dies is explained.
Figure 5. Clustering energy vs number of nodes .
Figure 6. Rounds un til th e first nod es d ie vs. nu mb er of nodes.
Figure 7. Rounds until the las t node dies vs. nu mber of nodes.

M. DHIVYA ET AL. 254
In Figure 8, the average energy consumed per rounds
is given. Number of alive nodes versus number of itera-
tions is given in Figure 9. Percentage of alive nodes ex-
plains the lifetime of the network. The number of itera-
tions is taken as 100. In Figure 10, the number of nodes
versus the percentage of nodes as cluster heads is given.
Normally the number of clusters is desirable to be less,
as it will help in transition of node states.
The energy consumption per node increases, as the
number of nodes increases. This is due to the fact that the
exchange of information between neighborhood nodes
and the radio channels to compete in the spectrum. The
number of cluster heads is high in LEACH and HEED
when compared to CBPA. As the node density increases,
depending on the application there are chances for the
number of clusters to increase or decrease in the tradi-
tional approaches. But in the proposed technique, it is
evident that the cluster head Percentage is within the
range.
Figure 8. Average energy consumed per round vs number of
rounds.
Figure 9. Number of alive nodes vs number of iterations.
Figure 10. Number of nodes vs % of nodes as cluster heads.
6. Conclusions
The cuckoo Based Particle Approach is developed to
achieve energy efficient Wireless Sensor Network and
multimodal objective functions. In this paper cuckoo
search is applied for cluster head selection and formation
of clusters among the Sensor nodes. The proposed CBPA
is compared with the standard LEACH protocol and
HEED protocol. The simulation results exhibits that
CBPA produces comparable results mainly due to opti-
mal search process in cluster formation and allocation of
appropriate paths in transmission of sensed data. The
developed suboptimal algorithm reduces complexity in
chain formation and prolongs the longevity of the Sensor
Network. The results are obtained by running more
number of simulations. The hybrid approach offers con-
sistency in the cluster formation, minimal number of clus-
ters, average energy consumption and energy consump-
tion per rounds.
In future, multi objective constraints are to be consid-
ered to obtain a realistic communication environment,
with scaling and system complexity. Hybrid Optimiza-
tion techniques combined with cross-layer design and
Machine/Parameter learning is a challenging issue in
research arena.
7. References
[1] F. Akyildiz, W. Su, W. Sankarasubramaniam and E. Cayirci,
“A Survey on Sensor Networks,” IEEE Communication
Magazine, Vol. 40, No. 8, August 2002, pp. 102-114.
[2] F. P. Ferentinos, T. A. Tsilgiridis, “Adaptive Design Op-
timization of Wireless Sensor Networks Using Genetic
Algorithms,” Computer Networks, Vol. 51, No. 4, 2007,
pp. 1031- 1051.
[3] M. Dhivya, M. Sundarambal and L. N. Anand, “A Re-
view of Energy Efficient Protocols for Wireless Sensor

M. DHIVYA ET AL. 255
Networks,” Proceedings of 1st International Conference
on Modeling, Control, Automation and Communication,
Tamilnadu, 20-21 December 2010, pp. 273-278.
[4] D. Wei, “Clustering Algorithms for Sensor Networks and
Mobile Ad Hoc Networks to Improve Energy Effi-
ciency,” Ph.D. Thesis, University of Cape Town, Ronde-
bosch, September 2007.
[5] A. G. Akojwar and R. M. Patrikar, “Improving Life Time
of Wireless Sensor Networks Using Neural Network
Based Classification Techniques with Cooperative Routing,”
International Journal of Communications, Vol. 2, No. 1,
2008, pp.75-86.
[6] T. F. Shih, “Particle Swarm Optimization Algorithm for
Energy-Efficient Cluster-Based Sensor Networks,” IEICE
Transactions on Fundamentals, Vol. E89-A, No. 7, 2006,
pp. 1950- 1958.
[7] Q. Zhao and L. Tong, “Energy-Efficient Information
Retrieval for Correlated Source Reconstruction in Sensor
Networks,” IEEE Transactions on Wireless Communica-
tions, Vol. 6, No. 1, 2007, pp. 157-165.
doi:10.1109/TWC.2007.04885
[8] W. Heinzelman, A. Chandrakasan and H. Balakrishnan,
“An Application-Specific Protocol Architecture for Wire-
less Microsensor Networks,” IEEE Transactions on Wire-
less Communications, Vol. 1, No. 4, 2002, pp. 660-670.
[9] N. Aslam, S. Sivakumar, W. Phillips and W. Robertson,
“Energy Efficient Cluster Formation Using a Mul-
ti-Criterion Optimization Technique for Wireless Sensor
Networks,” Proceedings of IEEE International Confer-
ence on Consumer Communications and Networking, Las
Vegas, 11-13 January 2007, pp. 650-654.
[10] O. Younis and S. Fahmy, “Heed: A Hybrid, Energy- Ef-
ficient, Distributed Clustering Approach for Ad Hoc
Sensor Networks,” Transactions on Mobile Computing,
Vol. 3, No. 4, 2004, pp. 660-669.
[11] X.-S. Yang and S. Deb, “Cuckoo Search via Levy Fli-
ghts,” Proceedings of World Congress on Nature & Bio-
logically Inspired Computing, New Delhi, 9-11 Dec-
ember 2009, pp. 210-214.
[12] D. X. Shuai, X. Wang and R Gong, “A Generalized Par-
ticle Model for Social Coordination and Autonomy in
MAS,” Proceedings of the 2nd IEEE International Con-
ference on Services Systems and Services Management,
Chongqing, 13-15 June 2005, pp. 985-990.
[13] X. Feng, F. C. M. Lau and D. X. Shuai, “A New Gener-
alized Particle Approach to Parallel Bandwidth Alloca-
tion,” Computer Communications, Vol. 29, No. 18, 2006,
pp. 3933-3945.
[14] D. X. Shuai, Q. Shuai, Y. M. Dong and L. J. Huang,
“Problem-Solving in Multi-Agent Systems: A Novel Gen-
eralized Particle Model,” Proceedings of the 1st IEEE
International Multi-Symposiums on Computer and Com-
putational Sciences, Hangzhou, 20-24 June 2006, pp.
322-329.
[15] D. X. Shuai and X. Feng, “Distributed Problem Solving
in Multiagent Systems: A Spring Net Approach,” IEEE
Intelligent Systems, Vol. 20, No. 4, 2005, pp. 66-74.
[16] D. X. Shuai, B. Zhang, C. P. Lu, “A New Generalized
Particle Dynamics Model for Software Cybernetics,”
Proceedings of IEEE International Computer Software
and Applications Conference, Chicago, 17-21 September
2006, pp. 240-245.
[17] A. Chakraborty, K. Chakraborty, S. K. Mitra and M. K.
Naskar, “An Energy Efficient Scheme for Data Gathering
in Wireless Sensor Networks Using Particle Swarm Op-
timization,” Journal of Applied Computer Science, Vol. 6,
No. 3, 2009, pp. 9-13.