Genetic algorithm based clustering approach for wireless sensor network to optimize routing techniques

Full text

Genetic Algorithm Based Clustering Approach for
Wireless Sensor Network to Optimize Routing
Techniques
Padmalaya Nayak
Department of IT, GRIET
Hyderabad, India
drpadmalaya2010@gmail.com
Bhavani Vathasavai
Department of CSE, MRCET
Hyderabad, India
bhavani91@gmail.com
Abstract— Since last decade, our eye witnessed proofs that
Wireless Sensor Networks (WSNs) have been used in many areas
like health care, agriculture, defense, military, disaster hit areas
and so on. The key parameters that play a major role in designing
a protocol for WSNs are its energy efficiency and computational
feasibility, as sensor nodes are resource constrained. Variation in
sensor nodes distance from base station and inter node distances
primarily cause unequal energy consumption among the sensor
nodes. The energy consumption varies with time and causes
degradation of system performance. LEACH is the first ever
clustered based routing protocol which provides elegant solutions,
suffers from the drawback due to the randomized cluster head
(CH) election. Assuming serious energy rebalancing with
traditional clustering algorithm, a Genetic Algorithm (GA) based
clustering algorithm which evaluates the fitness function by
considering the two major parameters distance and energy has
been proposed in this paper. GA is a probabilistic search based
algorithm based on the principle of natural selection and
evolution. Simulation result proofs that the proposed protocol
performs better than LEACH protocol and enhances the network
lifetime.
Index Terms— WSN, Clustering, Genetic Algorithm
I. I
NTRODUCTION
Wireless Sensor Network considered as real time
embedded system deployed in harsh environments where direct
human access is not possible. The nature of streamed data,
limited resources, distributed nature of WSNs brings many
more challenges in data routing since communication is often
more expensive than any other paradigm. Further, the
application under the data communication paradigm must be
energy efficient. In WSNs, many routing protocols are designed
based on the applications. For instance, the requirement of a fire
fighter differs w.r.t soil monitoring system. Fire fighter may
require timely updates for knowing the present fire conditions
whereas soil monitoring system requires the report periodically.
Hence, designers of routing protocols must consider the nature
of sensor nodes, types of application, and requirements of the
architecture etc. Some routing protocols do not satisfy the
quality of Service (QoS) required by real time applications even
if these are proven to be energy efficient. There are few
objectives that required keeping in mind while designing a
routing protocol.
These are:
• Energy consumption: It is universally acknowledged that
the routing protocol must be energy efficient so that
substantial lifetime of WSN can be accomplished.
• Load balancing: Energy consumption must be balanced
among all the sensor nodes so that lifetime of the WSN
can be prolonged.
• Clustering techniques: Decentralized equal number of
clustering techniques must be adopted to make the work
fast and simple.
• Scalability: When the network grows in size, excessive
communication overhead must not be increased even if it
is unavoidable to construct the path to the sink.
• Robustness: A reliable and robust protocol is always
desirable for efficient delivery of data in a proper time.
• Real time Application: Timely communication by
reducing end to end delay is an added feature of data
aggregation. Normally, timeliness of data with data
aggregation never go together.
Many energy efficient routing protocols based on various
techniques are discussed in the literature to make the network
alive for longer period of time [1-14]. Clustering is one of these
efficient techniques, where sensor nodes are grouped to form a
cluster and the cluster is managed by the Cluster Head (CH).
The CH gathers the data, compress it and send the compressed
data to the BS. The general system model for clustered WSN is
depicted in Fig.1. LEACH [1, 2] is the first famous hierarchical
routing protocol and seems to be most efficient over traditional
flat routing protocol. LEACH considers the probabilistic model
to elect the CH and tries to balance the load at each sensor node
in a rotation basis. Even though LEACH claims the efficiency
of the protocol, still there are some shadow areas that are
discussed in section II. In this paper, we have made attempt to
introduce a novel concept such as Genetic Algorithm (GA) to
elect an efficient CH which can balance the network load
among the sensor nodes and tries to find a suitable path to the
base station. GA is an evolutionary algorithm inspired from
biology. Many optimization problems have been solved by
using GA based algorithm. It considers three major operators
like fitness function, crossover and mutation. The protocol is
discussed in detail in section V.
376
978-1-5090-3519-9/17/$31.00 c
2017 IEEE

Fig. 1. An Example of clustered WSN
The rest of the paper is structured as follows. Section II
outlines the existing work. Section III presents the Radio
Model. Section IV discusses the proposed algorithm. Section
V discusses the simulation results followed by a conclusive
remark in Sec VI.
II. R
ELATED
W
ORK
In this section, most of the famous hierarchical routing
protocols based on the probabilistic models are discussed. We
have also discussed on some GA based routing protocols.
Although many clustered based protocols are discussed in the
current literature, very few well known protocols are discussed
here based on the interest of our work.
A. Hierarchical Routing Protocols based on clustering
LEACH [1,2] is an example of hierarchical routing
protocol which elects CH based on probabilistic model and
assuming that each sensor node has equal chance to become a
CH. This protocol operates in two phases; set up phase and
steady state phase. In set up phase, cluster formation takes place
and in steady state phase, actual data is transmitted. Each node
chooses a random number between 0 and 1 to become the CH.
If the number is less than the threshold value T(n), the node gets
the chance to be the cluster head for the current round. The
threshold value T(n) is defined in equation (1).
ܶሺ݊ሻൌ൝
௣
ଵି௣כቀ௥௠௢ௗ
భ
೛
ቁ
ǡ݂݅݊אܩ
Ͳǡ݋ݐ݄݁ݎݓ݅ݏ݁ (1)
r is the round which already ended, p is the probability of
the nodes to be the CH, G is a set of nodes which have never
been cluster head in the last 1/p rounds.
Although LEACH protocol distributes the load equally on
each cluster head, still there are some pitfalls that are discussed
here. There is no guarantee that preferred number of CHs is
elected in each round. If the elected node is located near to the
boundary of the network, other nodes could dissipate more
energy to transfer the message to CH. In LEACH-C [2], CH is
elected by the BS using a centralized algorithm. BS knows the
location information and energy of each node. So, it can
produce better clusters by dispersing Cluster Head nodes
throughout the network. In [3-11], many clustering protocols
have been discussed. SPIN [3] is a data gathering protocol,
which considers data negotiation to reduce data redundancy and
save energy. Directed diffusion [4] is another protocol and its
variant proof to be efficient by consuming less energy. MCEN
[5], SMECN [6], GAF [7], GEAR [8] rely on the accuracy of
the location information to optimize the energy conservation.
In [9], an energy aware QoS-routing protocol is discussed that
finds a least cost energy efficient path while meeting certain
end-to-end delay constraints. SPEED [10], as another QoS
routing protocol, depends on state-less geographical
information. Sequential Assignment Routing (SAR) [11] is the
first routing protocol for sensor networks that creates multiple
trees routed from one-hop neighbors of the sink by taking into
consideration both energy resources.
B. Genetic Algorithm based Routing Protocol
Science 1990s, GA has been used in many domains to solve
optimization problems and operates on Darwinian principle of
the survival of the fittest evolutionary theory. It has appealing
advantages over traditional mathematical programming based
algorithms in terms of handling complex problems. GA is based
on genetics and it has been successfully applied for solving
diverse optimization problems, including Multi Constraints
Path Problems (MCPS) and Multi Constraints Optimal Paths
(MCOPs) of WSNs. GA uses the fitness function to evaluate
the chromosomes, controls the evolutionary process and
produces the output. Motivated by this fact, we have applied
GA on clustering to optimize the routing techniques which can
extend network lifetime. Many researchers discuss many
different energy efficient routing protocols by using geographic
information, data aggregation, clustering approach etc. A
number of papers have demonstrated the usefulness of a GA
based approach in sensor networks [12], [13] [14], [15].
A Genetic Algorithm based approach for routing in two-
tiered sensor networks is proposed in [12] that only used energy
optimization to maximize life time of network. The work of
[14] focuses on finding an optimal traffic distribution to
improve the lifetime of multi-sensor networks. The work of
[15] discusses on deriving an energy efficient scheme that
satisfies the required detection probability using a distributed
GA. In [16], the author has reduced the no. of transmissions
between nodes in order to gain minimum energy. In [17], the
author has used Genetic Algorithm for routing in WSN in the
perspective of power management and energy communication
technique. The work discussed in [18] focuses on the
optimization on communication distance to build an energy
efficient routing for each sensor node. The author proposed a
routing algorithm based on multi objective algorithm [19].
However, most of the protocols are application specific. Here,
we have proposed the clustering protocol that evaluates the
fitness function by considering the two important factors such
as distance and energy.
2017 7th International Conference on Cloud Computing, Data Science & Engineering – Confluence 377

III. E
NERGY
M
ODEL
A
NALYSIS
Fig. 2. Radio Model
Each cluster head is the agent for the respective cluster to
collect the information from other nodes and send to the base
station either directly or indirectly. Since the amount of energy
consumption is proportional to the transmission and reception
of information, energy optimization model is considered and
shown in Fig. 2. Over a distance d, to transmit l bits of data
requires the amount of energy consumption from the transmitter
to the receiver is given in equation 2.
ܧ
்௫
ሺ݈ǡ݀ሻൌܧ
்௫ି௘௟௘௖
ሺ݈ሻ൅ܧ
்௫ି௔௠௣
ሺ݈Ǥ݀ሻ
ൌቊ݈כܧ
௘௟௘௖
൅݈כߝ
௙௦
כ݀
ଶ
݂݅݀൏݀
଴
Ǣ
݈כܧ
௘௟௘௖
൅݈כߝ
௠௣
כ݀
ସ
݂݅݀൒݀
଴
Ǣ (2)
• E
elec
implies the energy dissipated per bit to run the
transmitter or the receiver circuit. It depends on the factors
like digital coding, modulation, filtering and spreading of
the signal.
•
ߝ
௙௦
Ƭߝ
௠௣
are the characteristics of the transmitter
amplifier where
ߝ
௙௦
is used for free space and
ߝ
௠௣
for
multipath.
As the distance between transmitter and receiver is less than
the threshold value d
0
, the free space model (d
2
power loss) is
used. Otherwise, the multipath fading channel model (d
4
power
loss) is used. Power control can be used to invert this loss by
appropriately adjusting the power amplifier. The equation 3
shows the amount of energy consumption to receive l bit of data
and equation 4 represents the threshold value d
0
, which is the ratio
of ߝ
௙௦
and ߝ
௠௣
. ܧ
ோ௫
ሺ݈ሻൌܧ
௘௟௘௖
כ݈ (3)
݀
଴
ൌඥߝ
௙௦
ߝ
௠௣
Τ (4)
IV. P
ROPOSED
A
LGORITHM
In our Proposed Model, a new routing strategy inspired by
genetic algorithm has been proposed to find a suitable path for
a sensor node to send the data to the BS. GA has been proven
theoretically and empirically to be a robust search technique.
Each individual in the GA provides a possible solution to the
problem. Further, each is associated with a fitness function into
the new generation of the population, using Darwinian’s
principle of the survival of the fittest. Specifically, Routing
optimization technique aims that each sensor node should find
a feasible path to send the data to the base station. This fact
motivated us to propose a clustering algorithm where CH is
elected based on an objective function. The major goal of GA
is to search for the globally near-optimal solutions by
repeatedly calculating the fitness function using exploration
and exploitation methods. GA is inspired by the bio-inspired
processes of initialization, evaluation, selection, crossover,
mutation, and replacement as depicted in Fig. 3.
A. System Assumption
In the proposed model, sensor nodes are considered to be
deployed densely/randomly to monitor the environment
continuously. The proposed model is shown in Fig. 4. All the
sensor nodes cannot move after the deployment including base
station
1) Each sensor node contains unique ID id
i
after deployment
2) Homogeneous networks have been considered such that
all the sensor nodes have initial equal energy.
3) The number of sensor nodes depends upon the size of the
applications. For easy description, we have assumed N no.
of sensor nodes
4) The distance between the node and the base station can be
computed based on the received signal strength
5) The transmission range of all the sensor nodes is identical.
Let R
c
be the radius of each sensor node and the distance
between two sensor nodes (U, V) is d
min
. Only if d
min
< R
c
,
sensor node U can communicate with the sensor node V.
6) The base station has infinite power with significant
transmit power.
B. Network Model
The proposed clustering algorithm is built up on top of the
LEACH. It is assumed that sensor nodes send the data only if
an event occurs. CH gathers these data, aggregates it and
transmits it to the base station. The total time period is divided
into number of rounds. In each round, cluster is formed. In basic
LEACH [2], the cluster formation algorithm also implies that
the no. of cluster per round is k, where k is a system parameter.
The optimal value of k (k optimal) in LEACH can be
determined analytically by computation and communication
energy model. For instance, if there are N nodes distributed
randomly over M×M region, and k clusters are assumed, then
there are N/k nodes per cluster (one CH and (N/k)-1) non-CH
nodes. In LEACH, in each clustering round, every node
generates a random number between 0 and 1. If the random
number for a particular node is less than the threshold value T,
the node gets the chance to become the CH. Minimization of
energy consumption and maximization of network lifetime is
the ultimate aim of any routing algorithm. To minimize the
energy consumption, here the idea is to elect the farthest node
with higher energy as the source node. The source node is the
first Cluster Head. Each node calculates the distance to the
source node as well as to the Base Station. Further, each node
calculates its own energy and total residual energy of all the
nodes in the cluster. The ratio of distance of each node to the
Cluster Head and the total distance (from sensor node to CH
and CH to base station) is defined as the part of the fitness
function. The Cluster Head is elected on a fitness function given
in equation 5.
378 2017 7th International Conference on Cloud Computing, Data Science & Engineering – Confluence

Fig. 3. Flow Chart of Genetic Algorithm
The fitness function is calculated on account of the cost which
considers both the distance and the remaining energy power.
The cost value of distance and energy is given in equation 6 and
7. The CH can send the data to other CH by finding a suitable
path as shown in Fig. 4.
C. Proposed Algorithm
Consider a WSN Network G (V, E) where V is the set of
nodes and E is the set of links, 
ȁܰȁൌ݊, can be specified set of minimal and distinct
communication patterns P = {P
1
, P
2
,..P
m
}, where each traffic
pattern P
k
= (S
k
, D
k)
has the sender node S
k
and destination node
D
k
, 1 k  m, where D
k
is the proper subset of all other nodes,
it is noted that m  n. Placement of CH nodes to communicate
with each other is shown in Fig. 5. The connectivity matrix of
the proposed model is given in Fig. 6. If there is a direct link
between two nodes available, it is replaced with 1 or it is
replaced as 0.
D. Terminologies Used
• Individual Population: Represents that a path from
source to destination
• Child: Implies a adjacent node corresponding to a node
• Chromosome: Implies a complete path from source to
Base Station
• Gene: Represents a string of adjacent nodes
• Chromosome Structure: Tends to all the possible path
from source to the destination
Inputs: A set of N Sensor Nodes along with location
Information.
Processing: Formation of Clusters and Cluster Head Selection.
Outputs: An ordered sequence of routes to deliver the message
to BS
E. Steps for Genetic Algorithm
Step 1: Generation of Initial Population: Initial population
can be generated through the source Nodes.
Step 2: Calculate Fitness Function: The fitness function can
be calculated using the formula given in equation 5. The cost of
distance D
c
and energy E
c
is given in equation 6 and 7
respectively.
ܨ݅ݐ݊݁ݏݏܸ݈ܽݑ݁׷ܧ
௩
ൌܦ
௖
൅ܧ
௖
(5)
Where, D
c
is the cost of the distance and E
c
is the cost of the
energy. The value of the D
c
and E
c
can be calculated as given
in the equation
ܦ
௖
ൌ
ௗ
೔೓
ௗ
೔೓
ାௗ
೓ೞ
(6)
ܧ
௖
ൌ
ா
ೃ೔
ா
೅ೃ೔
(7)
Where, d
ih
is the distance of source sensor node to cluster head
and d
hs
is the distance of cluster head to base station. E
Ri
is the
energy of each sensor node and E
Tri
is the total residual energy
of all the sensor nodes within the cluster.
/* for every round */
1. Apply the connectivity of the Network to create the initial
gene population //repeat//
2. In each round, compute the value of fitness function
3. Number the paths in ascending order based on the value
of fitness function
4. If the fitness value of the first chromosome  1, then
terminate Else
5. Use Lower number paths to generate new genes
/* end of rounds */
Step 3: N times do;
3.1: Select Parents
3.2: Perform Crossover
3.3: Evaluate Offspring to select or reject
3.4: Perform Mutation
3.5: Repair selected offspring
3.5: Store the produced offspring for next generation
Step 4: Find the feasible path otherwise go to Step 2
Step 5: Stopping Criteria: The genetic operations are applied
until a route with fitness value is found i.e till the energy is
completely depleted.
2017 7th International Conference on Cloud Computing, Data Science & Engineering – Confluence 379

Fig. 4: Proposed Model of WSN
Fig. 5. Example Network with CH nodes
Fig. 6. Connectivity Matrix of Example Network
Fig. 7. Path 1through neighbor node 5 (Cr1)
Fig. 8. Path 2 through neighbor node 3 (Cr2)
Fig. 9. Path 3 through neighbor node 2 (Cr3)
Fig. 10. Path 4 after crossover Cr1 and Cr2
(New chromosome Cr4)
380 2017 7th International Conference on Cloud Computing, Data Science & Engineering – Confluence

Fig 11. Path 5 after Cross over Cr1 Cr2
(New Chromosome Cr5)
Fig. 12: After Mutation in Chromosome 2 (Cr2)
TABLE 1. INITIAL POPULATION WITH FITNESS VALUE
Chromosome No. Initial Population Fitness Value
Chromosome 1 (Cr 1) 1-5-3-6-7 4.86
Chromosome 2 (Cr 2) 1-3-5-4-7 3.41
Chromosome 3 (Cr3) 1-2-5-3-6-7 2.89
Chromosome 4 (Cr 4) 1-5-4-7 1.99
Chromosome 5 (Cr 5) 1-3-6-7 0.98
F. Explanation of the Algorithm
The proposed algorithm is demonstrated with 7 nodes. All
the nodes are replacement of CHs. The initial chromosome
structure is shown in Fig. 7, Fig. 8, and Fig. 9. The initial
population has been created by using step 1. The initial
population with chromosome value and fitness function is
calculated using the equation 5 as shown in Table 1.
Chromosome 1 (Cr
1
) and Chromosome 2 (Cr
2
) is elected as the
parents and crossover operation is performed. The crossover
point is CH
3
and CH
5.
After performing the crossover operations
on Cr1 and Cr2 (Fig. 7 and Fig. 8), the Path 4 (Fig. 10) and Path
5 (Fig. 11) is found. Chromosome 4 (Cr
4
) and chromosome 5
(Cr
5
) is added in the Table 1. In Fig.12, the mutation operation
is performed. After mutating the chromosome 2 at the mutating
point CH
3
, the resulting offspring found is same as chromosome
4 (Cr
4
).
V. S
IMULATION
R
ESULTS AND
A
NALYSIS
The validity of the proposed protocol has been verified
through NetSim simulator. It is a professional tool used to
measure the performance metrics of any type of protocol. The
results have been compared with LEACH protocol. The
simulation results proofs the efficiency of the proposed protocol
by finding a shortest path from source to destination. While
doing so, it consumes less energy and extends the network
lifetime.
A. Experimental Set-Up
For experimental set up, the proposed protocol considers
50 nodes deployed randomly over the area of (x=0, y=0) and
(x=100, y=100) with base station location (x=50, y=50). The
proposed algorithm assumes four no. of clusters. Each round
duration is 20s. The bandwidth of the channel is 1 Mbps. Each
data message is 500 bytes long. Packet header length is 25
bytes. We have used a simple energy model. The
communication parameters and the required parameters of
interest are given in Table 2. We run the simulation for 20000s.
After running the simulation extensively, it is concluded that
the proposed approach performs better than LEACH which is
discussed in the next section.
TABLE 2. SIMULATION PARAMETERS
Type
Parameters
Value
Network Topology
Network Size
No. of Nodes
Expected no. of Clusters
BS Location
Node distribution
BS Mobility
Channel
Channel type
100x100m
40
5
50x50m
Random
Random Walk
Wireless
Bidirectional
Radio Model
Energy Model
Start Up Energy
ETx elec/ERx
׫fs
׫mp
Battery
2J
50nJ/bit
10pJ/bit/m2
0.0013pJ/bit/m4
Application
Simulation time
Round time
Packet Header Size
Data Packet Size
Bandwidth
20000s
20s
25 bytes
500 bytes
1Mbps
2017 7th International Conference on Cloud Computing, Data Science & Engineering – Confluence 381

B. Results and Discussion
In this section, the experimental results obtained from the
simulation tool are discussed to measure the performance of
the proposed algorithm. Fig. 13 shows that the first node dies
in 77.5 seconds in LEACH-MH whereas it survives more than
88 seconds in GA-LEACH justifying that the life time of the
whole network depends on the lifetime of individual sensor
node. Fig. 14 conveys that more number of data packets
(including data and control packet) is transmitted in GA-
LEACH compared to LEACH. Fig. 15 shows that the number
of data signals delivered to BS is more in GE-LEACH
compared to LEACH. It is universally accepted that Network
lifetime can be extended to the optimum value by minimizing
the energy consumption as much as possible. To verify this, we
have checked the survival of sensor nodes w.r.t number of
rounds. Fig. 16 proofs that in GA-LEACH, sensor nodes
survive up to more number of rounds compared to LEACH.
Running the simulations for longer period of time, it is
confirmed that when only 40 nodes alive in LEACH w.r.t.
25000 rounds, 90 nodes alive in the purposed algorithm.
Fig. 13. First node dies over time
Fig. 14. Total no. of packets transmitted
Fig. 15. Total no. of data signals delivered to BS
Fig. 16. Network lifetime over time
VI.
C
ONCLUSION
LEACH is a promising protocol and provides basic building
blocks to develop many more protocols on top of it. In this
paper, an energy efficient clustering algorithm has been
proposed for Wireless Sensor Network to optimize the network
lifetime based on Genetic Algorithm concept. Genetic
Algorithm is a natural bio inspired algorithm solves many
optimization problems. By electing suitable fitness function,
Cluster Head is elected in each round. Each Cluster Head sends
the data to the neighboring Cluster Head in a chain like system
to compute a suitable a path till it reaches at the Base Station.
Simulation results confirms that the proposed algorithm
performs better than LEACH and resumes longer network
lifetime.
R
EFERENCES
[1] W. R. Heinzelman, A. Chandrakasan, and H. Balakrishnan, Energy-
efficient Communication Protocol for Wireless Microsensor Networks.,
in IEEE Computer Society Proceedings of the Thirty Third Hawaii
International Conference on System Sciences (HICSS '00), Washington,
DC: IEEE, 2000, pp. 1-10.
382 2017 7th International Conference on Cloud Computing, Data Science & Engineering – Confluence

[2] W. R. Heinzelman, A. Chandrakasan and H. Balakrishnan,“An
application-specific protocol architecture for wireless microsensor
networks,” in IEEE Transactions on Wireless Communications, 1(4),
pp. 660 – 670, Oct. 2002.
[3] W.R Heinzelman, Kulki J, Balakrishna H.,Adaptive protocols for
information dissemination Wireless Sensor Networks”, in Proc. Of
MOBICOM 99’ Seattle, WA, 1999.
[4] Intangonwiwat C, Govinda R, Estrin D, “Directed diffusion: a scalable
and robust paradigm for Sensor Networks”, in Proc. Of ACM
MOBICOM’00, Boston, MA, 2000.
[5] Rodoupolu V, Meng TH., “Minimum energy mobile wireless
networks”,IEEE Journal Selected Areas in Communications 1999, 17,
1333–1344.
[6] LilL, HALparn JY, “Minimum energy mobile wireless networks
revisited”, In Proc. Of ICC Conference, Helsinki, Finland, 2001.
[7] Xu Y,. Heidemann J, Estrin D, “Geography-informed energy
conservation for ad hoc routing”, In Proc. Of the 7
th
Annual ACM/IEEE
International Conference on Mobile Computing and Networking, Rome,
Italy, 2001.
[8] YU Y, Estrin D, Govindan R, “Geographical and energy aware routing:
a recursive data dissemination protocol for Wireless Sensor Networks”,
UCLA Computer Science Department Technical Report, TR-01-0023,
May 2001.
[9] Gao Q, Blow KJ, Holding DJ, Marshall I, Peng XH, “Radio range
adjustment for energy efficient wireless sensor networks”, Ad hoc
Networks 2006, Vol. 4(1), 75-82.
[10] Hea T, Stankovia JA, Lub C, Abdelzahera T, “SPEED: a stateless
protocol for real-time communication in sensor networks”, 23
rd
IEEE
International Conference on Distributed Computing System
(ICDCS’03), Rhoda Island, USA, 2003
[11] T.Abdlzaher, T.He, J. stankonvic, “Feedback control of data aggregation
in sensor networks, in Decision and Control, 2004.
[12] Atail. Bari, ShamsulWazed, Arunita Jackle, Subir Bandyopadhyay, “ A
Genetic Algorithm based approach for energy efficient routing in two
tired Sensor Networks” Ad hoc Networks, Vol. 17., 2009, 665-676
[13] Navrati Saxsena, Abhisek Roy, Jitai Shen, “ A QoS based Energy
Efficient Sensor Routing Protocol” Wireless Communication and
Mobile Computing Journal, 2009, Vol. 9,No.3, 417-426.
[14] Y. Pan, X. Liu, “Energy Efficient lifetime maximization and sleeping
schedule data fusion and QoS Multi SensorNet” Signal Processing,
2007, Vol. 87 (12), 2949-2964.
[15] Q.Qiu, et. al, “Lifetime aware and Resource Management for sensor
Network using distributed Genetic Algorithm” ACM Press, New York,
2006, 191-196.
[16] Jin, M.zhou, A-S Wu, “Sensor Network Optimization using Genetic
Algorithm” in Proc. Of the 7
th
World Multiconference in Systemics
Cybernetics and Informatics, 2003.
[17] Babie. S, Ad Khadim, “A new method for improving lifetime in WSN
by using Genetic Algorithm” Advances in Intelligent and Soft
Computing, Vol. 30, No. 8, 2012, pp. 177-184.
[18] AI-obaidy. M, et.all, “Optimizing the communication distance of an ad
hoc Sensor Networks by Genetic Algorithm” Artificial Intelligence
Review, Vol. 29, No. 3-4, PP. 183-194, 2008.
[19] Rong, J.L, Xin, X.X and Tian, X.S., “Energy balancing algorithm in
Wireless Sensor Network”, Journal of Benjing Universities of
Technology, Vol. 38, No. 5, 2012, pp. 740-743.
[20] Padmalaya Nayak, D. Anurag, “ A Fuzzy Logic based Clustering
Algorithm to Extend the Network Lifetime” IEEE Sensor Journal, Vol.
16. No.1, 2016, PP. 137-144.
[21] D.V Puspalatha, Padmalaya Nayak,” A Clustering Algorithm for WSN
to optimize the Network Lifetime using Type 2 Fuzzy Logic” in Proc.
Of 3
rd
International Conference on Artificial Intelligent, Modelling and
Simulations (AIMS 2015), UMS, Malaysia, PP. 53-58
2017 7th International Conference on Cloud Computing, Data Science & Engineering – Confluence 383