CFGA : Clustering wireless sensor network using
fuzzy logic and genetic algorithm
Esmaeil saeedian
Islamic Azad University of Mashhad
Iran
e_saeedian@yahoo.com
Mehrdad Jalali
Islamic Azad University of Mashhad
Iran
mehrjalali@gmail.com
Mohammad Mahdi Tajari
Islamic Azad University of Mashhad
Iran
mm.tajari@gmail.com
Massoud niazi Torshiz
Islamic Azad University of Mashhad
Iran
masood.niazi@gmail.com
Ghamarnaz Tadayon
Islamic Azad University of Mashhad
Iran
gh_tadayon@yahoo.com
Abstract - Wireless sensor networks have numerous nodes with
limited energy which have the ability to monitor around
themselves and these nodes are scattered in a limited
geographic area. One of the important issues in these networks
is increasing the network lifetime. In this study, we introduce
an efficient protocol for trade-off between loads and increase
in life expectancy of the network, known as CFGA (Clustered
wsn using fuzzy logic and genetic algorithm) which uses the
single - step method for intracluster communication and multi
- step for intercluster communication. At the beginning of
each round, each node first checks its fuzzy module, and based
on output of the fuzzy module, if the clusterhead capability
exists for node, it would be ready ( in each region, number
of the best nodes would be ready), then at the base station by
using genetic algorithm and the location of clusterheads based
on minimum energy consumed, the optimum network nodes
are determined. In this study, consumption of energy for nodes
which are not capable of becoming clustered is prevented and
also only nodes with high capabilities in genetic algorithm take
part in order to converge faster to the optimum solution.
Keywords- Genetic Algorithm; balance energy; wireless sensor
networks; cluster classification;fuzzy logic
I. INTRODUCTION
In the recent years, technological advances and the
telecommunications industry, electrical and electronic micro
components, leading to construction of small and relatively
inexpensive sensors which relate to each other through
wireless communication [1].
The networks that to be known wireless sensor networks,
have become to a suitable tool for extracting data from the
environment and monitoring environmental events and their
applications in household, industrial and military, is
increasing day to day [2].
In design and development of wireless sensor networks
the main issue is sensors energy source limitation. Because
of too many sensors in the network, or lack of access to
them, replacement or charging sensors’ batteries are not
practical. For this reason providing ways to optimize energy
consumption, which ultimately increases the network life
time, is of great importance[1].
Research has shown that network nodes organized into
groups called clusters, can be more efficient in reducing
energy consumption, which leads to increase the network
lifetime. The network lifetime is the interval running
protocol up to the first dead in Nodes [3].
The LEACH protocol [1] is one of the propounded
clustering protocols in sensor networks which are made of
two setup phase and steady state phase.
In steady state phase data transfer is done in a single hop
manner to the sink. In each cluster one node is selected as a
cluster head and the rest of node’s clusters are called cluster-
member. Data collected from member nodes are processed in
cluster head before sending to a base station or sink, and then
will be sent to the base station after aggregation of data in the
form of a package.
Since cluster head’s energy consumption is more than
ordinary nodes, after a while their energy will be over.
Therefore, in LEACH, using dynamic clustering has been
introduced. This means that after each send and receive
course operations, in other words after each launch of
implementation phase, cluster heads are changed randomly
and another node is chosen as cluster head. But random
selection of cluster heads may lead to improper distribution
in the network. This means that in a part of the network
under coverage, two or more adjacent cluster heads are
placed close together, whereas in any other region there
would not be a cluster head.
978-1-4244-6252-0/11/$26.00 ©2011 IEEE