CLUSTERING IN WIRELESS SENSOR NETWORK BASED ON GENETIC ALGORITHM

Abstract

The article considers the solution of an important scientific and practical task of increasing
the lifetime of a wireless sensor network (WSN) as part of distributed information systems. Wireless sensor
networks are one of the integral components of modern distributed information systems which represent
packet data networks, united by a set of locally spaced intelligent sensor devices consisting of a microcontroller, a set of sensors (data collector), battery and transceiver module. Such networks are widely used in
environmental monitoring, in security systems, etc. and serve to obtain the required information (for example, temperature, humidity, seismic data, etc.), which is then transmitted to the base station for further processing. The main parameter of the wireless sensor network is the network lifetime, which is largely determined by energy resources. The sensor network must have a sufficient lifetime to meet the task, for example,
several months or several years. Restrictions on energy resources lead to the fact that the network should
assume low power consumption. To solve the problem of power consumption, routing protocols using various network topologies are being developed. To increase the wireless sensor network lifetime, an improved
approach to the formation of a clustered network structure that combines the advantages of a genetic algorithm performed according to a conventional scheme and k-means method used as part of initial population
in a genetic algorithm is proposed. This approach allows increasing the speed of the clustering algorithm,
making it less dependent on the initial data. To reduce intracluster and intercluster distances, the Davies–
Bouldin index is used as a fitness function. Based on the proposed clustering method, the KGACVI protocol
was developed. The simulation results show that the developed KGACVI protocol using the proposed clustering algorithm showed better results than the compared protocols (SEP, IHCR and ERP) when comparing the
network lifetime with different number of heterogeneous nodes due to the decrease in power consumption for
data transmission from node to base station.